MARINE ECOSYSTEMS, FISHERIES AND CLIMATE CHANGE

Over the last few years, several studies have discussed the potential impacts of climate change on marine ecosystems and marine wild capture fisheries. The IPCC Fourth Assessment Report (Parry et al. Reference Parry, Canziani, Palutikof, Parry, Canziani, Palutikof, van der Linden and Hanson2007) concluded that climate change was affecting marine ecosystems and the production of fish, and was likely to confound the impacts of natural variation on fishing activities and to complicate management efforts. These changes were occurring because of altered food web dynamics, decreased abundances of benthic habitat-forming species, shifting species distributions and a greater incidence of fish diseases (Parry et al. Reference Parry, Canziani, Palutikof, Parry, Canziani, Palutikof, van der Linden and Hanson2007; Hoegh-Guldberg & Bruno Reference Hoegh-Guldberg and Bruno2010). However, there are also likely to be significant regional variations to these global patterns (Table 1). For example, whereas increased vertical stability is likely to reduce nutrient inputs to large areas of the ocean, at higher latitudes this increased stability should also increase the residence times of plankton within the well-lit euphotic zone, thereby increasing their productivity.

Table 1. Climate-related threats and potential consequences to marine ecosystems based on spatial scale. Modified from Brierley & Kingsford (Reference Brierley and Kingsford2009)

Barange & Perry (Reference Barange, Perry, Cochrane, De Young, Soto and Bahri2009) provided a detailed overview of scientific knowledge on the physical and ecological impacts of climate change on marine and inland capture fisheries and aquaculture based on the literature up to 2007. Warming of sea waters will have direct effects on the distributions and abundances of marine populations, with consequent changes in the production from fisheries. Distributions will shift poleward, and seasonal migrations are likely to occur earlier. Changes in fish abundances are expected, with populations at the polar extents of their ranges likely to increase in abundance, whereas those at their equatorial extents will decrease. Warmer temperatures will also increase growth rates although, in practice, these may be limited by the availability of suitable food. For example, changes in the species composition of marine zooplankton in the European North Sea have been suggested as an important contributing factor to declines in Atlantic cod (Gadus morhua) abundances in that area (Beaugrand et al. Reference Beaugrand, Edwards, Brander, Luczak and Ibanez2008). Changes in the timing of important ecological processes have been observed, for example changes in the plankton community structure in the European North Sea in which dinoflagellate blooms are peaking earlier as a result of warming whereas the larger diatom species show no consistent pattern of change.

Recent advances in modelling studies

Recently, there has been considerable progress in estimating the potential impacts of climate change on marine ecosystems, in particular with respect to modelling. Several approaches are now in use in order to forecast the effects of climate change on fish and fisheries (Hollowed et al. Reference Hollowed, Barange, Ito, Kim and Loeng2010). These include global static models, global dynamic models, dynamic downscaling approaches and statistical downscaling approaches. Hollowed et al. (Reference Hollowed, Barange, Ito, Kim and Loeng2010) noted that statistical time-series analyses can describe previous patterns of variability but may be less useful for forward projections. Cheung et al. (Reference Cheung, Close, Lam, Watson and Pauly2008, Reference Cheung, Lam, Sarmiento, Kearney, Watson and Pauly2009, Reference Cheung, Lam, Sarmiento, Kearney, Watson, Zeller and Pauly2010) have made significant quantitative advances on projecting the impacts of climate change on distributions of fish populations, commercial fish and invertebrate biodiversity, and fisheries catch potential using a statistical bioclimatic envelope approach (based on the ranges of temperature and other physical conditions within which species occur combined with trophic energetics and allometric scaling of metabolism; Cheung et al. Reference Cheung, Close, Lam, Watson and Pauly2008). Using high (IPCC A1B), medium (IPCC B1) and low (stable, 2000) emissions scenarios, they projected range shifts of 45–60 km per decade, with 0·80 of species moving towards the poles. By 2050, average shifts were 600 km for pelagic species and 220 km for demersal (bottom fish) species (Cheung et al. Reference Cheung, Lam, Sarmiento, Kearney, Watson and Pauly2009). The global average species extinction rate was 0·03, with the Arctic and Southern Ocean experiencing the highest (5·5 and 2 times the global average, respectively) and the equatorial region the lowest (0·5 times the global average). Cheung et al. (Reference Cheung, Lam, Sarmiento, Kearney, Watson and Pauly2009) noted this is lower than the predicted extinction rates of 0·15–0·37 for terrestrial systems, concluding that climate change is likely to result in numerous local extinctions of fish populations in sub-polar, tropical and semi-enclosed marine regions, and that species invasion will be most intense in polar regions. When projecting the implications of these results for the potential fisheries catch in ten-year averages from 2005 to 2055, Cheung et al. (Reference Cheung, Lam, Sarmiento, Kearney, Watson, Zeller and Pauly2010) found little overall change in global maximum catch potential (±1%) but considerable change in the spatial distribution of this potential catch. Under the high-emissions scenario, catch potential in higher-latitude regions increased on average by 30–70% whereas it decreased by up to 40% in the tropics (e.g. Fig. 1). The Indo-Pacific was projected to be most highly impacted, with up to a 50% decline in ten-year average maximum fisheries catch potential by 2055. Other regions with projected large losses were Indonesia, the continental United States, Chile and China. The largest increases in catch potential were projected to occur off Norway, Greenland, Alaska and eastern Russia. Under the low-emissions scenario, however, the pattern of changes was less clear, being about 0·60 of the results from the high-emissions scenario.

Fig. 1. Model-based projected changes in ten-year averaged maximum catch potential from 2005 to 2055 by the 20 exclusive economic zone regions with the highest catch in the 2000s. The numbers in parentheses represent the numbers of exploited species included in the analysis. Reprinted from Cheung et al. (Reference Cheung, Lam, Sarmiento, Kearney, Watson, Zeller and Pauly2010) with permission of Wiley-Blackwell Publishing.

Models of climate change impacts to ocean primary production have had variable results, with some projecting slight decreases in global average production (e.g. Cox et al. Reference Cox, Betts, Jones, Spall and Totterdell2000) and others slight increases (e.g. 0·7–8·1%; Sarmiento et al. Reference Sarmiento, Slater, Barber, Bopp, Doney, Hirst, Kleypas, Matear, Mikolajewicz, Monfray, Soldatov, Spall and Stouffer2004), although with large spatial variability. For the western North Pacific Ocean, results from a 3D primary production model using a more complex food web and full coupling of the biology and the physics project a slight decrease in annual average chlorophyll biomass due to increased vertical stratification, but, more importantly for fish, a shift towards earlier timing and a change in the dominant phytoplankton species (from large diatoms to smaller forms) of the spring bloom (Hashioka & Yamanaka Reference Hashioka and Yamanaka2007).

Simulation models have now progressed to the stage such that they can be coupled directly with atmospheric models of climate change to project impacts on fish populations. Brown et al. (Reference Brown, Fulton, Hobday, Matear, Possingham, Bulman, Christensen, Forrest, Gehrke, Gribble, Griffiths, Lozano-Montes, Martin, Metcalf, Okey, Watson and Richardson2010) used the common marine food web model Ecopath with Ecosim (EwE) linked with an ocean general circulation model (CSIRO Mark 3.5 coupled atmosphere–ocean general circulation model) and a nutrient–phytoplankton–zooplankton lower trophic level model to evaluate the impacts of climate change on primary production and fisheries around Australia. Under a standard climate change scenario (IPCC A2), they projected increases in primary production, increases in fisheries catch and value and increases in the biomass of threatened marine species such as turtles and sharks. Brown et al. (Reference Brown, Fulton, Hobday, Matear, Possingham, Bulman, Christensen, Forrest, Gehrke, Gribble, Griffiths, Lozano-Montes, Martin, Metcalf, Okey, Watson and Richardson2010) also projected little change in species composition using this model. However, model formulations that included more complex predator and competition interactions reversed the expected responses from some species resulting in declines in catches and in abundances of some threatened species even with background increases in primary production.

The United Kingdom QUEST_FISH program (Barange et al. Reference Barange, Allen, Allison, Badjeck, Blanchard, Drakeford, Dulvy, Harle, Holmes, Holt, Jennings, Lowe, Merino, Mullon, Pilling, Rodwell, Tompkins, Werner, Ommer, Perry, Cury and Cochranein press a) is developing coupled physical to fish to people models to project the implications of global change on marine ecosystems. They couple global circulation models to high-resolution physical–biological models of 20 large marine ecosystems, and then estimate size-based fish production from primary production and temperature using macro-ecological theory (e.g. Jennings et al. Reference Jennings, Mélin, Blanchard, Forster, Dulvy and Wilson2008). This approach permits the assessment of future climate changes combined with non-climate influences such as global fish commodity markets. Initial results are indicating that the sustainability of the global fishmeal market and its supporting fisheries depends as much (or more) on how society responds to climate impacts as on the strength of the climate effect (Merino et al. Reference Merino, Barange and Mullon2010a).

Recent advances in observational studies

Observational evidence continues to accumulate that suggests the impacts of climate change are occuring now and are not just theoretical future concerns, although distinguishing these impacts from climate variability and the effects of direct human stresses such as fishing can be difficult (e.g. Rouyer et al. Reference Rouyer, Fromentin, Ménard, Cazelles, Briand, Pianet, Planque and Stenseth2008; Last et al. Reference Last, White, Gledhill, Hobday, Brown, Edgar and Peclin press). Many of these observed changes are helping in verifying the projections of climate change models. Some of the most evident climate-related changes are occurring in polar regions. Sea ice is particularly sensitive to climate change, with predictions that the Arctic Ocean may be ice free in summer by 2030 (Brierley & Kingsford Reference Brierley and Kingsford2009), yet sea ice is the single most important physical feature structuring the marine ecosystems in these regions. Schofield et al. (Reference Schofield, Ducklow, Martinson, Meredith, Moline and Fraser2010) described how rapid environmental change in the West Antarctic Peninsula region over the past few decades has coincided with shifts in the marine food web from plankton to apex predators, significantly altering marine ecosystem structure and function. Physical changes include increases in mid-winter surface air temperatures of 6°C (five times faster than the global average) over the past 50 years, 0·87 of the glaciers in this region are in retreat, and the loss of perennial sea ice. Biological changes include a 12% decrease in the magnitude of phytoplankton blooms over the past 30 years, a net decrease in primary production and changes in species composition towards smaller-celled organisms. As a result, the keystone species Antarctic krill (Euphausia superba) is being replaced in this region by soft-bodied tunicates as the important zooplankton grazers, which has caused changes in top predators. Abundances of the ice-dependent Adélie penguin (Pygoscelis adeliae) have declined by 90%, whereas more northerly ice-intolerant penguin species have invaded the area and are increasing in abundance. These changes can be clearly related to climate change since fishing does not play a role in this particular regional ecosystem.

Climate change is also clearly impacting on coral reef systems. Mass coral bleaching and resulting mortalities due to increasing temperatures are already reducing the density and diversity of coral reef fishes and other organisms (Hoegh-Guldberg et al. Reference Hoegh-Guldberg, Mumby, Hooten, Steneck, Greenfield, Gomez, Harvell, Sale, Edwards, Caldeira, Knowlton, Eakin, Iglesias-Prieto, Muthiga, Bradbury, Dubi and Hatziolos2007). Coral bleaching occurs as temperatures exceed 1°C above the long-term summer maximum for extended periods (Hoegh-Guldberg et al. Reference Hoegh-Guldberg, Hughes, Anthony, Caldeira, Hatziolos and Kleypas2009). Munday et al. (Reference Munday, Jones, Pratchett and Williams2008) suggested that such bleaching will have rapid impacts on the diversity and species composition of coral reef fish communities. At present, however, there is little evidence that coral bleaching had has much impact on coral reef fisheries, perhaps because these fisheries do not target the smaller and specialized coral-dependent species that are more vulnerable to environmental change (Munday et al. Reference Munday, Jones, Pratchett and Williams2008; Pratchett et al. Reference Pratchett, Munday, Wilson, Graham, Cinner, Bellwood, Jones, Polunin and Mcclanahan2008). Climate change is likely to have additional impacts on the trophic linkages, recruitment dynamics and population connectivity of coral reef fishes (Munday et al. Reference Munday, Jones, Pratchett and Williams2008; Table 2).

Table 2. Predicted impacts of climate change on populations and communities of coral reef systems, and the level of uncertainty associated with these predictions (high: predications are well supported by observations; moderate: predictions have some empirical support but are not conclusive; low: predictions are speculative and lack good supporting evidence). Modified from Munday et al. (Reference Munday, Jones, Pratchett and Williams2008)

For the North-eastern Atlantic, Rijnsdorp et al. (Reference Rijnsdorp, Peck, Engelhard, Möllmann and Pinnegar2009) concluded that global warming is principally responsible for large-scale shifts in the distribution and abundances of fish species in this region. Changes in the distributions of demersal species have generally been smaller than changes in the distributions of pelagic species. Rijnsdorp et al. (Reference Rijnsdorp, Peck, Engelhard, Möllmann and Pinnegar2009) suggested that climate-related changes in recruitment success, in particular increased survival of the early life stages, and improvements to the nursery habitats for juvenile fishes, were key factors in these changes. A deepening of the distribution of the bottom fish community in the European North Sea as bottom temperatures warmed has also been observed (Dulvy et al. Reference Dulvy, Rogers, Jennings, Stelzenmüller, Dye and Skjoldal2008), indicating that distribution shifts will occur in three dimensions. Hiddink & Ter Hofstede (Reference Hiddink and Ter Hofstede2008) examined changes in the diversity of fish species in the North Sea over the past 22 years and found that increases in the richness of species were related to increasing water temperatures, and that much of this increase was due to smaller-sized species mostly of southerly origin. Pinnegar et al. (Reference Pinnegar, Cheung and Heath2010) reviewed the potential impacts of climate change to marine fisheries around the British Isles (Table 3) and concluded that fishing locations for traditional commercial species are changing, which is affecting the effectiveness of location-based management measures. In addition, new fisheries are being developed for warm-water species which now occur regularly in UK waters. Potential future changes not only include slightly higher (1–2%) fishery yields by 2050, but also the disappearance of some traditional fishery stocks by the end of the century.

Table 3. Summary of observed and future predicted changes to marine fisheries in the waters around the British Isles (from Pinnegar et al. Reference Pinnegar, Cheung and Heath2010)

A PICES Scientific Report edited by Beamish (Reference Beamish2008) examined whether climate trends affect the production of species in the major fisheries of the North Pacific and found that the specific impacts depended on the life history of each species. It concluded that, although the mechanisms linking climate trends to fisheries production in this region are still poorly understood, the amount of prey available to first feeding young was likely to be a common constraint. As the Bering Sea has warmed and sea ice has retreated, there have been shifts in species distributions as well as overall increases in total biomass, species richness and average trophic level (Mueter & Litzow Reference Mueter and Litzow2008). However, fish species did not all shift distributions at the same rates. This caused traditional fish communities to be pulled apart and to be reorganized into new communities (Mueter & Litzow Reference Mueter and Litzow2008), potentially affecting fisheries and ecosystem productivity.

There is observational evidence that the production of plankton, which forms the base of marine food webs leading to commercial fisheries, has been changing globally. Boyce et al. (Reference Boyce, Lewis and Worm2010), using data on ocean transparency and chlorophyll biomass since 1899, reported declines of phytoplankton biomass in eight out of ten ocean regions, at an estimated rate of about 1% of the global median per year. The largest rates of decline were in the South and Equatorial Atlantic and in the Southern Ocean. Increasing biomass was observed in the northern and southern Indian Ocean. Boyce et al. (Reference Boyce, Lewis and Worm2010) concluded that these long-term declining trends were related to increasing sea surface temperatures. Using a 9-year time series of remotely sensed ocean colour data, Polovina et al. (Reference Polovina, Howell and Abecassis2008) found that the world's least productive waters (mostly associated with open ocean gyres) have expanded by 15% from 1998 to 2006, consistent with increasing sea temperatures but at rates greater than projected by model studies. O'Connor et al. (Reference O'Connor, Piehler, Leech, Anton and Bruno2009) tested experimentally the effects of warmer temperatures on a simplified marine food web (consisting of phytoplankton and zooplankton). As temperatures were increased, it was found that grazing by zooplankton increased at faster rates than increases in the primary productivity of the phytoplankton, with the net result that standing phytoplankton biomass was reduced. The end result was a decline of the overall biomass in their experimental system. They concluded that even small changes in water temperature can cause significant shifts in marine food web structure and productivity. In addition, Morán et al. (Reference Morán, López-Urrutia, Calvo-Diaz and Li2010) found that warmer ocean temperatures lead to smaller phytoplankton cells, further altering the structure of marine ecosystems. Richardson (Reference Richardson2008) suggested that global warming may have greater impacts on marine than terrestrial ecosystems because temperature can have important influences on the abundance, size composition, diversity and trophic efficiency of the zooplankton. Impacts of global warming on zooplankton have been observed as poleward shifts in distributions and earlier timing of life history events such as spawning, leading to changes in abundance and community structure. It was concluded that range shifts of zooplankton are among the fastest and largest of any marine or terrestrial group.

The potential impact of ocean acidification on marine systems, both on its own and in combination with other stressors such as warming temperatures, is a rapidly progressing research topic (Doney et al. Reference Doney, Fabry, Feely and Kleypas2009). Huge unknowns remain, largely due to the lack of basic knowledge of biological and chemical responses of marine organisms to low pH. Organisms with calcium carbonate shells or skeletons will certainly be negatively impacted. Cooley & Doney (Reference Cooley and Doney2009) provided a study of the potential impacts of ocean acidification on fisheries, in particular those based on molluscs, in the United States and suggested that a reduction of 0·1–0·2 pH units over the next 50 years would result in harvest declines of 6–25% in mollusc fisheries, for an estimated loss of several billion US dollars. It was concluded that damage to marine habitats and resources due to ocean acidification would cause substantial declines in fisheries revenues and jobs, and have other indirect economic costs related to ecosystem impacts on fish species which are predators of these invertebrate species. In the UK, Pinnegar et al. (Reference Pinnegar, Cheung and Heath2010) noted that four of the ten most valuable marine fisheries are those for calcifying invertebrates (Nephrops, scallops, crabs and lobsters), which contribute 0·44 of the total value from fisheries. Evidently, there is the potential for significant economic harm to fisheries and marine food security due to ocean acidification.

MANAGEMENT AND SOCIETAL IMPLICATIONS

Models are beginning to be developed which project the impacts of global warming from marine ecosystems to fisheries production and their consequences for human societies. These are, however, ‘early days’ for these approaches, with many uncertainties over how information is transformed within human communities and how human behaviour responds to changing pressures (Hollowed et al. Reference Hollowed, Barange, Ito, Kim and Loeng2010). Allison et al. (Reference Allison, Perry, Badjeck, Adger, Brown, Conway, Halls, Pilling, Reynolds, Andrew and Dulvy2009) assessed the vulnerability of national economies to the potential impacts of global warming on fisheries using an approach in which vulnerability was based on the exposure to the physical effects of global warming, the sensitivity (dependence) of the national economy on fisheries and the extent to which the nation had the adaptive capacity to mitigate the potential impacts. It was observed that many of the most vulnerable countries were also among the least developed and with a high dependence on fish in local diets. These included four countries in West and Central Africa, Peru and Columbia in north-western South America, and four countries in tropical Asia. It is noteworthy that most of these countries are in the equatorial region, a region which the modelling studies of Cheung et al. (Reference Cheung, Lam, Sarmiento, Kearney, Watson and Pauly2009, Reference Cheung, Lam, Sarmiento, Kearney, Watson, Zeller and Pauly2010) predicted will be impacted significantly and negatively by the effects of global warming on fish diversity and fisheries catch potential. Merino et al. (Reference Merino, Barange and Mullon2010a, Reference Merino, Barange, Mullon and Rodwell2010b) modelled the marine production–human consumption system of small pelagic fisheries and fishmeal markets under climate variability and change. It was concluded that regional fish stocks were able to recover from climate fluctuations unless these occurred at the same time as increasing fishing pressure due to expansions of international markets. When the model system was forced with a climate change scenario, the results suggested that the sustainability of small pelagic fish populations depended more on how society responded to the climate impacts than on the magnitude of the climate changes (Barange et al. Reference Barange, Cheung, Merino and Perryin press b).

A number of responses and adaptation options are available for human societies to respond to the impacts of climate change on marine ecosystems and fisheries production (Daw et al. Reference Daw, Adger, Brown, Badjeck, Cochrane, De Young, Soto and Bahri2009; Badjeck et al. Reference Badjeck, Allison, Halls and Dulvey2010). These include changing fisheries operations, switching target species, alternative fishing occupations, short-term bridging strategies and ultimately the abandonment of fishing. McIlgorm et al. (Reference Mcilgorm, Hanna, Knapp, Lefloc'h, Millerd and Pan2010) noted that fishery governance will need to deal with uncertainty in both the ecological system and in the human social system. It was suggested that adaptive governance systems will need to include more flexible fishery management approaches (which are able to respond more rapidly to changing conditions), schemes for adjusting fishing capacity to new situations, catch limitations and development of alternative livelihoods for fishers. Grafton (Reference Grafton2010) explored the problem of when, and how, marine capture fisheries should adapt to climate changes and proposed that a risk and vulnerability assessment and management decision-making framework be developed prior to the actual need to adapt. Such a framework would include an assessment of current and potential future vulnerabilities, engage stakeholders and simulate a variety of response options to different future conditions. Without such a decision framework in place before major changes occur, Grafton (Reference Grafton2010) noted that current actions risk increasing vulnerabilities or being inadequate to the challenges of climate change and recognized two complementary and interacting strategies: developing measures that increase the resilience and adaptive capacities of marine ecological and human social systems before significant impacts occur, and adaptive management actions to quickly respond to unexpected changes when they are observed to occur. The economics of adapting fisheries to climate change has been considered by a number of recent activities (e.g. Hannesson et al. Reference Hannesson, Barange and Herrick2006; Hannesson Reference Hannesson2007), most recently by the Organisation for Economic Cooperation and Development at a workshop in June 2010 (Workshop on the Economics of Adapting Fisheries to Climate Change; www.oecd.org). A report on the potential costs of adapting to a 2°C warmer world by 2050, prepared for the World Bank (World Bank 2009), estimated global losses in landed catch value of US$10–31 billion by 2050, with countries in East Asia and the Pacific being the most severely impacted. Losses for developing countries globally ranged from US$7 to US$19 billion, whereas losses for developed countries ranged from US$2 to US$8 billion. Change in Europe was estimated to range from a loss of US$1 billion to a slight gain.

Specific examples of management strategies to address issues of climate change and fisheries are beginning to appear in the scientific literature. Higgason & Brown (Reference Higgason and Brown2009) presented a framework that includes workshops with stakeholders to discuss the potential local impacts of climate change, research plans to address key uncertainties and an approach to adaptively manage changes as they occur (Fig. 2); the application of this framework to the Gulf of the Farallones National Marine Sanctuary in California was presented. Stram & Evans (Reference Stram and Evans2009) described how the United States North Pacific Fisheries Management Council is developing risk-averse management approaches in advance of increasing uncertainty due to climate change. Pro-active measures taken by the council have included closing US arctic waters to all commercial fishing pending further research, assessing whether changes in fish distributions might allow for unregulated fishing in ‘new’ locations, and adopting measures to control the expansion of the trawl fishing fleet into newly ice-free areas. In addition, adaptive management measures are being developed for those species for which the links between climate variables and fish distributions have been identified.

Fig. 2. Process to manage the effects of climate change on a local ecosystem, as developed for the Gulf of the Farallones National Marine Sanctuary, California. Redrawn from Higgason & Brown (Reference Higgason and Brown2009), by permission of Oxford University Press.

KNOWLEDGE GAPS AND NEXT STEPS

The impact of climate change on marine ecosystems, fisheries and food security is a rapidly progressing field of inquiry, and great progress has been made over the past few years, in particular with respect to modelling. An international symposium on The Effects of Climate Change on Fish and Fisheries was held in Sendai, Japan, in April 2010 to discuss recent advances and current work on this topic. A preliminary summary of the key outcomes from the meeting is provided by Hollowed et al. (Reference Hollowed, Barange, Ito, Kim and Loeng2010; Table 4). Two key outcomes from this symposium are highlighted: the need for long-term and co-ordinated ocean monitoring to observe the impacts of climate changes, and further improvements to modelling capabilities.

Table 4. Summary of preliminary key outcomes from the Symposium on Effects of Climate Change on Fish and Fisheries, held 26–29 April 2010 in Sendai, Japan. Modified after Hollowed et al. (Reference Hollowed, Barange, Ito, Kim and Loeng2010)

A co-ordinated approach to observational studies is needed to separate the impacts of climate change from those of climate variability. With the latter, there is an expectation that conditions will eventually return to ‘normal’, whereas with the former, ‘normal’ may have to be redefined. The response strategies of human communities may be very different between these two situations (Perry et al. Reference Perry, Ommer, Allison, Badjeck, Barange, Hamilton, Jarre, Quinones, Sumaila, Barange, Field, Harris, Hofmann, Perry and Werner2010a). A strategy for an observational system to detect climate change impacts in the West Antarctic Peninsula region would involve a nested multi-platform approach, including deployment of oceanographic sensors on marine animals (Schofield et al. Reference Schofield, Ducklow, Martinson, Meredith, Moline and Fraser2010). The goals of such a system would be to quantify the atmosphere and ocean heat budgets, to understand how the deep ocean interacts with the Antarctic shelf region, and how these affect the regional climate, ice dynamics and marine ecology (Schofield et al. Reference Schofield, Ducklow, Martinson, Meredith, Moline and Fraser2010). Beaugrand et al. (Reference Beaugrand, Edwards, Brander, Luczak and Ibanez2008) pointed out that some regions of the ocean may be more important than others and have greater leverage on the impacts of climate change and identified a critical thermal boundary in the North Atlantic (between 9 and 10°C which coincides with the transition zone between two large marine biomes), at which small changes in temperature can produce large ecosystem shifts across several trophic levels, including Atlantic cod (Gadus morhua). Such locations are obvious sites for enhanced monitoring, and other high-leverage locations need to be identified. These global observational systems then need to be connected to a network with an efficient data distribution system so that comparative analyses can be done to separate climate change from variability (e.g. IOC 2009).

Although modelling the impacts of climate change on marine ecosystems and fisheries has made significant advances, several important challenges remain. Brander (Reference Brander2009) suggested the application of five criteria to judge the quality of predictions of climate change impacts: (1) do they provide additional information beyond that derived from first principles; (2) are their predictions testable; (3) are they credible and provide confidence levels and uncertainties relating to confounding factors; (4) are they capable of predicting past changes; and (5) are they based on knowledge of physical, physiological and ecological structures, processes and limits? For example, the bioclimatic envelope approach to climate impacts projections has been criticized as being too simple, and that physiological responses to stress in nature are more complex. Instead, a systems-based approach is recommended to integrate the effects of environmental and physiological variability on the ecological responses of marine organisms (Helmuth Reference Helmuth2009). As noted by Brown et al. (Reference Brown, Fulton, Hobday, Matear, Possingham, Bulman, Christensen, Forrest, Gehrke, Gribble, Griffiths, Lozano-Montes, Martin, Metcalf, Okey, Watson and Richardson2010), including more realistic species interactions such as predation and competition can significantly alter the projections of ecosystem models. Feedback effects, such as how changes in demographic characteristics, mortality rates and allele frequencies affect growth rates, organism size and resilience of the entire population to climate change (Brierley & Kingsford Reference Brierley and Kingsford2009), and the impacts of multiple stressors on marine systems have generally not yet been included in climate impacts models. Substantial research is needed on the direct and synergistic impacts of ocean acidification on fisheries production. In addition, most of these are not undisturbed populations and systems, and many have been fished for a long time. The activity of fishing, in particular industrial-scale fishing over the past 60 years, changes the characteristics of marine populations, communities and ecosystems (Planque et al. Reference Planque, Fromentin, Cury, Drinkwater, Jennings, Perry and Kifani2010) so that they no longer have the same adaptive capacities to climate variability and change and may respond more rapidly than in the past (Perry et al. Reference Perry, Cury, Brander, Jennings, Möllmann and Planque2010b). Genner et al. (Reference Genner, Sims, Southward, Budd, Masterson, Mchugh, Rendle, Southall, Wearmouth and Hawkins2010) examined the long-term (1911–2007) variability of a bottom fish assemblage in the western English Channel, a region experiencing both commercial fisheries and wide (2°C) interannual temperature changes; it was found that abundances of smaller-sized species followed the changes in temperature, suggesting they would likely respond quickly to climate changes. In contrast, the larger-sized species declined in both abundance and mean body size, as expected with sustained size-based overharvesting. Lindegren et al. (Reference Lindegren, Möllmann, Nielsen, Brander, Mackenzie and Stenseth2010) provided an example of the type of regional-scale modelling that is needed to assess the combined impacts of climate change and fishing on marine food webs, and the consequences for management actions. The results obtained for Atlantic cod in the Baltic Sea show how exploitation will need to be adjusted to achieve sustainable management with different climate change scenarios.

Urgent problems that need to be addressed regarding the impacts of climate change to marine wild capture fisheries include reducing the (presently large) uncertainties in climate change impacts models, understanding the synergistic effects of multiple stressors beyond changes in temperature and fishing, and inclusion of humans with improved realism as both drivers and recipients of changes in marine ecosystems. Stock et al. Reference Stock, Alexander, Bond, Brander, Cheung, Curchitser, Delworth, Dunne, Griffies, Haltuch, Hare, Hollowed, Lehodey, Levin, Link, Rose, Rykaczewski, Sarmiento, Stouffer, Schwing, Vecchi and Werner(in press) reviewed the application of IPCC-type climate models to living marine resources, and concluded that uncertainties in these models, relating to coarse resolution and complexities of marine ecosystem responses, can limit the robustness and precision of the projections from such models; analyses of multi-model ensembles, bias corrections and statistical and dynamic downscaling to smaller spatial scales to resolve some of these limitations were recommended. Regional models are better able to incorporate smaller-scale features that control primary production, and can also include other characteristics such as changes in phenology and species (size) composition of organisms at the base of the food web. Other stressors may also become apparent at regional scales, such as low oxygen (Weston et al. (Reference Weston, Fernand, Nicholls, Marca-Bell, Mills, Sivyer and Trimmer2008) provide an example for the North Sea), and these need to be included along with changes in temperature, salinity, pH and fishing. Assessments of the socio-economic impacts of climate change on marine capture fisheries should develop the vulnerability approach adopted by the IPCC (e.g. Allison et al. Reference Allison, Perry, Badjeck, Adger, Brown, Conway, Halls, Pilling, Reynolds, Andrew and Dulvy2009; Johnson & Welch Reference Johnson and Welch2010). This approach considers vulnerability as a function of exposure, sensitivity and adaptive capacity of the socio-economic system to climate changes. Such an approach also needs to be applied at regional and local spatial scales to get at local motivations and responses to change (e.g. Brookfield et al. Reference Brookfield, Gray and Hatchard2005). The ultimate problem is with the limits of predictability itself. If marine systems do develop ‘no-analogue’ communities as a result of shifting species distributions at different rates and other impacts, then understanding of, and the inferences from, previous conditions may not apply (Williams & Jackson Reference Williams and Jackson2007).