Future climate change and its effects on the ecosystem and human activities

Будущие перспективы
Typography

Over the last 50 years, air temperatures have increased almost twice as fast in the Arctic than the global average. Models predict that air temperatures will continue to increase considerably, and summer sea ice in the Arctic is likely to disappear before the middle of this century and winter sea ice by the end of the current century (IPCC, 2013). Because of the complex dynamics of the Barents Sea ecosystem, and because the effects of climate change will interact with other major factors, such as

acidification and the impact of fisheries, it is difficult to predict what the total effect on this ecosystem will be. However, it can be predicted with fair certainty that some of the ice-associated fauna and flora in the Barents Sea will be lost or at least significantly reduced. Also, a number of species, e.g. cod and capelin, will likely have a more northern and/or eastern distribution and boreal species such as blue whiting and mackerel may become common in the Barents Sea. These changes will likely result in potentially large changes in community composition, and it is possible that the structure of the ecosystem may shift irreversibly. The probability of this happening may increase if the pressures from other types of impacts, such as fisheries and acidification, are high.

Projections of future climate change – Global perspective

On the global level, the Intergovernmental Panel on Climate Change (IPCC) is the leading international body for the assessment of climate change.  IPCC’s most recent and Fifth Assessment Report (AR5) was released between September 2013 and November 2014. The main conclusion of AR5 is that “Warming of the climate system is unequivocal, and since the 1950s, many of the observed changes are unprecedented over decades to millennia. The atmosphere and ocean have warmed, the amounts of snow and ice have diminished, and sea level has risen”.

One of the main reasons for global warming is cumulative emissions of CO2 (IPCC, 2013). Projections of future anthropogenic greenhouse gas (GHG) emissions vary over a wide range depending on both socio-economic development (mainly driven by population size, economic activity, lifestyle, energy use, land use patterns, and technology) as well as climate policy. AR5 used Representative Concentration Pathways (RCPs) for developing various future scenarios of GHG levels.  They described four different 21st century pathways of GHG emissions and atmospheric concentrations, air pollutant emissions and land use. The RCPs include a stringent mitigation scenario (RCP2.6), two intermediate scenarios (RCP4.5 and RCP6.0), and one scenario with very high GHG emissions (RCP8.5). Scenarios without additional efforts to constrain emissions (’baseline scenarios’) led to pathways ranging between RCP6.0 and RCP8.5 (Figure 5.1.1). RCP2.6 is representative of a scenario that aims to keep global warming likely below 2°C above pre-industrial temperatures.

Figure 5.1.1. Annual emissions of CO2 in gt yr-1 showing the mean and the range for different RCPs.Figure 5.1.1. Annual emissions of CO2 in gt yr-1 showing the mean and the range for different RCPs.

Multiple lines of evidence indicate a strong, consistent, almost linear relationship between cumulative CO2 emissions and projected global temperature change to the year 2100 in both the RCPs and the wider set of mitigation scenarios.  Based upon the CO2 emission variants the AR5 concluded: “surface temperature is projected to rise over the 21st century under all assessed emission scenarios. It is very likely that heat waves will occur more often and last longer, and that extreme precipitation events will become more intense and frequent in many regions. The ocean will continue to warm and acidify, and global mean sea level to rise”.

Global mean surface air temperature change for the period 2016–2035 relative to 1986–2005 is similar for the four RCPs and will likely be in the range 0.3°C to 0.7°C (medium confidence) (Figure 5.1.2).  Assumptions in developing the simulations were that there will be no major volcanic eruptions or changes in some natural sources (e.g., CH4 and N2O), or unexpected changes in total solar irradiance. By mid-21st century, the magnitude of projected climate change is substantially affected by the choice of emissions scenario.

Increase of global mean surface temperature by the end of the 21st century (2081–2100) relative to 1986–2005 is likely to be 0.3°C to 1.7°C under RCP2.6, 1.1°C to 2.6°C under RCP4.5, 1.4°C to 3.1°C under RCP6.0 and 2.6°C to 4.8°C under RCP8.59. The AR5 conclude “the Arctic region will continue to warm more rapidly than the global mean”.

Figure 5.1.2. Global average surface temperature change (a) from 2006 to 2100 as determined by multi-model simulations. Time series of projections and a measure of uncertainty (shading) are shown for scenarios RCP2.6 (blue) and RCP8.5 (red). The
	mean and associated uncertainties averaged over 2081–2100 are given for all RCP scenarios as colored vertical bars at the right hand side of panel. The number of Coupled Model Intercomparison Project Phase 5 (CMIP5) models used to calculate the multi-model
	mean is indicated.Figure 5.1.2. Global average surface temperature change (a) from 2006 to 2100 as determined by multi-model simulations. Time series of projections and a measure of uncertainty (shading) are shown for scenarios RCP2.6 (blue) and RCP8.5 (red). The mean and associated uncertainties averaged over 2081–2100 are given for all RCP scenarios as colored vertical bars at the right hand side of panel. The number of Coupled Model Intercomparison Project Phase 5 (CMIP5) models used to calculate the multi-model mean is indicated.

Figure
	5.1.3. Annual mean zonal temperature change in the ocean for 2081-21000 relative to 1986–2005 under RCP2.6 (left), RCP4.5 (centre) and RCP8.5 (right) forcing scenarios. Hatching indicates regions where the multi-model mean change is less than one standard
	deviation of the internal variability. Stippling indicates regions where the multi-model change mean is greater than two standard deviations of the internal variability and where at least 90% of the models agree on the sign of change.Figure 5.1.3. Annual mean zonal temperature change in the ocean for 2081-21000 relative to 1986–2005 under RCP2.6 (left), RCP4.5 (centre) and RCP8.5 (right) forcing scenarios. Hatching indicates regions where the multi-model mean change is less than one standard deviation of the internal variability. Stippling indicates regions where the multi-model change mean is greater than two standard deviations of the internal variability and where at least 90% of the models agree on the sign of change.

Figure 5.1.3 shows multi-model mean projections of zonal-averaged ocean temperature change under three emission scenarios. Differences in projected ocean temperature changes for different RCPs become more distinct as the century progresses. The largest warming is found in the top few hundred meters of the subtropical gyres, similar to the observed pattern of ocean temperature changes (Levitus et al., 2012). Surface warming varies considerably between the emission scenarios ranging from about 1°C (RCP2.6) to more than 3°C in RCP8.5. Mixing and advection processes gradually transfer the additional heat to deeper levels of about 2000 m at the end of the 21st century. Depending on the emission scenario, global ocean warming between 0.5°C (RCP2.6) and 1.5°C (RCP8.5) will reach a depth of about 1 km by the end of the century. This figure also shows that at latitudes higher than 75°N for all scenarios water temperature will increase.

In the Technical Summary of WG1 for AR5 it was noted that Arctic sea-ice cover, very likely, will continue shrinking and thinning year-round during the course of the 21st century (Figure 5.1.4).

Figure 5.1.4. Northern Hemisphere (NH) sea-ice extent in September
	over the late 20th century and the whole 21st century for the scenarios RCP2.6, RCP4.5, RCP6.0 and RCP8.5 in the CMIP5 models, and corresponding maps of multi-model results in 2081–2100 of NH September sea-ice extent. In the time series, the number of
	CMIP5 models to calculate the multi-model mean is indicated (subset in brackets). Time series are given as 5-year running means. The projected mean sea-ice extent of a subset of models that most closely reproduces the climatologic mean state.  The
	1979–2012 trend of the Arctic sea ice is given (solid lines), with the minimum to maximum range of the subset indicated with shading. Black (grey shading) is the modeled historical evolution using historical reconstructed forcing. The CMIP5 multi-model
	mean is indicated with dashed lines. In the maps, the CMIP5 multi-model mean is given in white and the results for the subset in grey. Filled areas mark the averages over the 2081–2100 period, lines mark the sea ice extent averaged over the 1986–2005
	period. The observed sea-ice extent is given in pink as a time series and averaged over 1986–2005 as a pink line in the map.Figure 5.1.4. Northern Hemisphere (NH) sea-ice extent in September over the late 20th century and the whole 21st century for the scenarios RCP2.6, RCP4.5, RCP6.0 and RCP8.5 in the CMIP5 models, and corresponding maps of multi-model results in 2081–2100 of NH September sea-ice extent. In the time series, the number of CMIP5 models to calculate the multi-model mean is indicated (subset in brackets). Time series are given as 5-year running means. The projected mean sea-ice extent of a subset of models that most closely reproduces the climatologic mean state.  The 1979–2012 trend of the Arctic sea ice is given (solid lines), with the minimum to maximum range of the subset indicated with shading. Black (grey shading) is the modeled historical evolution using historical reconstructed forcing. The CMIP5 multi-model mean is indicated with dashed lines. In the maps, the CMIP5 multi-model mean is given in white and the results for the subset in grey. Filled areas mark the averages over the 2081–2100 period, lines mark the sea ice extent averaged over the 1986–2005 period. The observed sea-ice extent is given in pink as a time series and averaged over 1986–2005 as a pink line in the map.

The CMIP5 multi-model projections give average reductions in Arctic sea-ice extent for 2081–2100 compared to 1986–2005 ranging from 8% for RCP2.6 to 34% for RCP8.5 in February and from 43% for RCP2.6 to 94% for RCP8.5 in September (medium confidence). A nearly ice-free Arctic Ocean (sea-ice extent less than 106 km2 for at least five consecutive years) in September before mid-century is likely under RCP8.5 (medium confidence), based on an assessment of a subset of models that most closely reproduce the climatological mean state and 1979–2012 trend for Arctic sea-ice cover. Some climate projections exhibit 5- to 10-year periods of sharp summer Arctic sea-ice decline — even steeper than observed over the last decade — and it is likely that such instances of rapid ice loss will occur in the future. There is little evidence in global climate models of a tipping point (or critical threshold) in the transition from a perennially ice-covered to a seasonally ice-free Arctic Ocean beyond which further sea ice loss is unstoppable and irreversible.

Figure 5.1.5. The mean anomalies of the surface air temperature for 2011–2030 (а, d), 2041–2060 (b, e) and 2080–2099 (c, f) by the end of the 21st Century for summer (а, b, c) and winter (e, f, g). The simulations were based on an ensemble of 31
	CMIP5 models using RCP4.5 scenarios.Figure 5.1.5. The mean anomalies of the surface air temperature for 2011–2030 (а, d), 2041–2060 (b, e) and 2080–2099 (c, f) by the end of the 21st Century for summer (а, b, c) and winter (e, f, g). The simulations were based on an ensemble of 31 CMIP5 models using RCP4.5 scenarios.

Projections of future climate change – Barents Sea

Projections for future changing surface air temperature over Russia and marine areas including the Barents Sea (Figure 5.1.5) and precipitation (Figure 5.1.6) for 2011-2030, 2041-2060, and 2080-2099 appeared in the Second Assessment Report for Climate Change in Russia.  In the Barents Sea, surface air temperature will increase more in winter (Figure 5.1.5 d, e, f) than in summer (а, b, c). For winter, maximum anomalies are situated between Svalbard and northern Novaya Zemlya for all periods and gradually increase from 2011-2030 (6-70C) to 2080-2099 (11-120C). In summer, the maximum anomaly area for 2080-2099 is situated in the central part of the Barents Sea and near the southern part of Novaya Zemlya, and the anomalies gradually increase from 2011-2030 (1-20C) to 2080-2099 (4-50C).

Precipitation is also projected to increase, being higher in winter (Figure 5.1.6 d, e, f) than in summer (а, b, c). For winter, maximum precipitation anomalies are situated between Svalbard and the southern coast of Russia for all periods and gradually increase from 2011-2030 (20-25%) to 2080-2099 (40-45%). For summer, maximum anomalies are situated in the northern Barents Sea and gradually increase from 2011-2030 (5-10%) to 2080-2099 (15-20%).

Figure 5.1.6.  The mean
	precipitation anomaly (%) for 2011–2030 (а, d), 2041–2060 (b, e) and 2080–2099 (c, f) for summer (а, b, c) and winter (d, e, f) based on an ensemble of 31 CMIP5 models using RCP4.5 scenarios.Figure 5.1.6.  The mean precipitation anomaly (%) for 2011–2030 (а, d), 2041–2060 (b, e) and 2080–2099 (c, f) for summer (а, b, c) and winter (d, e, f) based on an ensemble of 31 CMIP5 models using RCP4.5 scenarios.

Projections of future physical oceanographic conditions in the Barents Sea

A number of early projections of oceanographic conditions for the Barents Sea have been made.  Furevik et al. (2002) suggested that by 2080, surface ocean temperatures will warm by 1° to 2°C, winter sea ice will almost disappear, Atlantic Water will spread farther eastward and northward, and the surface mixed-layer depth will increase due to stronger winds.  Ellingsen et al. (2008) suggested that sea-ice coverage will decrease with the largest decline in summer and virtually ice-free summer conditions by 2059. A 25% increase in freshwater runoff to the Barents Sea and a peak spring discharge 2-3 weeks earlier than at present was projected by Dankers and Middelkoop (2008). In spite of this, an increase in future salinity has been predicted owing to higher salinities in the Atlantic Water inflows caused by higher evaporation in the tropics (Betke et al., 2006). 

Huse and Ellingsen (2008) examined changes in the position of the Polar Front that separates the cold Arctic Water and warm Atlantic Water. The frontal position was projected not to change much in the western Barents, where it is tied to topographic features, but in the eastern Barents the front was projected to move farther north and east.  In a more recent study by Wassmann et al. (2015), using a Regional Circulation Model (RCM) called SINMOD, it is suggested that the front may move all the way to the northern shelf break adjacent to the Arctic Ocean (Figure 5.1.7).  This would result in much warmer waters, especially in the northern Barents Sea (Figure 5.1.7 A, C).

Figure 5.1.7. The average position of the Polar Front in April.  The
	frontal position is indicated by the boundary between the waters less than -1°C (blue) and above +1°C (red) in panels A and C.  Also shown in these panels is the average current vector at 50 m depth.  The gross primary production for April
	is shown in panels B and D. The years 2000-2009 are displayed in A and B while 2090-2099 are in C and D.Figure 5.1.7. The average position of the Polar Front in April.  The frontal position is indicated by the boundary between the waters less than -1°C (blue) and above +1°C (red) in panels A and C.  Also shown in these panels is the average current vector at 50 m depth.  The gross primary production for April is shown in panels B and D. The years 2000-2009 are displayed in A and B while 2090-2099 are in C and D.

Most of the future climate scenarios that have been developed for the Barents Sea are based on low resolution (order 100 kms) Global Circulation Models (GCMs).  More recently, Regional Climate Models (RCMs) have been developed with much higher spatial resolution (order 10s of kms).  Sandø et al. (2014) used two GCMs, the GISS Ocean-Atmosphere Model and the NCAR CCSM3, to downscale to a regional model of the Barents Sea based on ROMS (Regional Ocean Modelling System).  The two global models were chosen based on their performance to recreate sea ice conditions in the Barents Sea (Overland and Wang, 2007).  Downscaled results for present day conditions in the Barents Sea were closer to the observations than for the two GCMs and the differences between the downscaled results from the two models were less than the differences between the two GCMs.  However, future scenarios from the two downscaled models were significantly different.  Downscaling using the NCAR model resulted in much higher heat transport into the Barents Sea and water masses became less saline compared to using the GISS model.  The authors concluded that RCM results depend largely on the GCM chosen to downscale from; hence, exactly what will happen in the Barents Sea under climate change remains somewhat uncertain.  One approach to overcome dependency on a particular GCM would be to undertake the downscaling using several GCMs and then take an ensemble mean.  This should provide a better estimate while spread in the model results would indicate the uncertainty in the projections.  Also, there is a need to couple atmosphere and ocean for the regional models, which even in the recent modelling by Sandø et al. (2014) was not attempted. In a coupled model system, changes in ocean feedback to the atmosphere are incorporated; in an uncoupled system, there is no feedback.

It is clear however that under climate change, temperatures will rise in the Barents Sea somewhere between 2°-10°C, and that sea ice will be significantly reduced and may disappear all together.  Salinities are generally expected to decline due to increasing precipitation and higher fresh-water runoff from rivers.  Peak river runoff will occur earlier in the year.  The Polar Front will move to the northeast and there will be greater amounts of Atlantic Water and less Arctic Water. 

Recent studies have provided new insights regarding climate variability in the Barents Sea climate. Sando et al. (2014) used results from the NorESM1-M coupled climate model to show that the negative trend in sea-ice coverage reflects the major trend of heat transport through the Barents Sea Opening. They concluded that the ocean has a stronger direct impact on changing sea-ice coverage than does the atmosphere.  Smedsrud et al. (2013) similarly concluded that loss of ice cover in the Barents Sea is driven by increased transport of heat into the region with inflowing Atlantic Water. These authors also found that correlations — between the Barents Sea ice coverage and the North Atlantic Oscillation — are highly variable yet remain relatively low over extended periods of time.  Drinkwater et al. (2014) supported earlier findings of a strong Atlantic Multi-decadal Oscillation (AMO) like signal in Barents Sea temperatures and in Arctic sea-ice variability.  The AMO has a period of 60-80 years and is believed to be linked to changes in the Meridional Overturning Circulation. 

 
Future foodwebs in the Barents Sea under climate change

It is clear that the projected changes in ocean climate indicated above will have significant impacts on organisms in the Barents Sea.  A full discussion on expected biological responses to the physical changes is beyond the scope of the present report.  However, in the following we present some changes which may occur.

Plankton

In seasonally ice-covered areas where the ice will disappear or be reduced under climate change, annual primary production is projected to increase due to higher light levels and an extended growing season (Arrigo et al., 2008). Indeed, satellite imagery suggested a significant increase in net primary production in the Barents Sea between 1998 and 2006 (Mueter et al., 2009).  Future projections vary slightly in terms of the estimated change in overall primary production in the Barents Sea. Ellingsen et al. (2008) suggested an increase of about 8% over a 65 year-long period, mostly occurring in the eastern and northeastern regions. No significant trend was found by Slagstad et al. (2011) when they examined the response to increases of 2° to 8°C in waters of the Barents Sea.  Skaret et al. (2014), on the other hand, projected an increase of 36% by 2046-2064 compared to 1981-1999, mostly in northern and eastern regions.  The varying results are due to differences in model formulation and the greenhouse gas scenarios used.  Moreover, not only primary production that will be impacted; ice algal communities will be lost or reduced as the sea ice declines.  Also, earlier phytoplankton blooms are likely under a warming climate scenario through earlier onset of density stratification of the water column (Ji et al., 2013).

Earlier sea ice melt and the subsequent release of ice algal communities into the water column at a time when surface waters are cold and zooplankton growth rates are low could result in low zooplankton abundance and reduced grazing, thereby increasing the sinking flux of particulate matter from the sea ice to the sediments (Arrigo et al., 2008).  Leu et al. (2011) found evidence that a mismatch between the timing of ice break-up and the growth period for Calanus glacialis caused up to a 5-fold decrease in their biomass.  If advection of warm surface waters is primarily responsible for early loss of sea ice, zooplankton growth may not be negatively impacted and carbon export may remain unchanged or even diminish (Arrigo et al., 2008). Reduced sea-ice cover has been proposed to favour a pelagic-dominated ecosystem over the more typical sea-ice algae to benthos ecosystem (Piepenburg, 2005). Such an ecosystem switch would reduce the vertical export of organic carbon and decrease pelagic-benthic coupling, despite an overall increase in phytoplankton productivity.
Ellingsen et al. (2008) predicted that under climate change Atlantic zooplankton production, primarily Calanus finmarchicus, would increase by about 20% and spread farther eastward while the Arctic zooplankton biomass would decrease significantly (by 50%) resulting in an overall decrease in zooplankton production in the Barents Sea.  The increase in Atlantic species biomass did not compensate for the losses of Arctic zooplankton resulting in an overall decrease of zooplankton biomass in the Barents Sea. The increased Atlantic zooplankton is caused by both higher transport into the Barents through inflow of warm Atlantic Water (Stenevik and Sundby, 2007) and to faster turnover rates due to the higher temperatures, as suggested by Tittensor et al. (2003) for the Labrador Sea.  Indeed, the loss of Arctic zooplankton species the northern Barents Sea has been observed during recent years in association with warm temperatures and reduced ice cover (Dalpadado et al., 2012). 

Fish and Shellfish

If warmer temperatures and sea ice reductions result in higher phytoplankton production in the Barents Sea, it is expected to result in increased fish production. For example, model studies suggest that higher primary production tends to increase cod recruitment in the Barents Sea (Svendsen et al., 2007). Drinkwater (2005) modelled the response of cod recruitment throughout the North Atlantic to future warming scenarios based upon previous responses to temperature variability. In the Barents Sea, he found increases in cod recruitment compared to present values under temperature increases of 1° to 4°C. Coupled with expected higher growth rates, there would be an increase in total cod biomass in the Barents Sea, which in turn could lead to increased fish catches (Drinkwater, 2005; Stenevik and Sundby, 2007). However, any increase in cod recruitment and abundance will depend on the changes in secondary (zooplankton) production, in particular C. Finmarchicus — the primary prey for cod larvae. The expected increased abundance of C. finmarchicus in the Barents Sea under future climate change (Ellingsen et al., 2008), supports the contention that cod recruitment will likely increase. However, not all models agree.  Kristiansen et al. (2014) predicted that cod production would decline later during the 21st century owing to a decrease in zooplankton productivity.

A northward and eastward shift in the distribution of cod under climate change was suggested earlier (Loeng et al., 2005; Drinkwater, 2005; Stenevik and Sundby, 2007; Cheung et al., 2008), with the potential for penetrating as far as the Kara Sea (Drinkwater, 2005). Such shifts have already been observed with cod reaching a historic latitudinal maximum in the Barents Sea at the continental shelf break adjacent to the Arctic Ocean (Kjesbu et al., 2014).  In addition, more cod spawning is expected to take place in the north and less in the southern regions along the coast of Norway (Stenevik and Sundby, 2007; Sundby and Nakken, 2008). Distribution shifts of adults will result in a higher proportion of cod in Russian waters, although because of expected increases in total production, numbers of fish in both the Norwegian and Russian economic zones are expected to increase.  This should result in increased fish catches by both Norway and Russia under climate change (Loeng et al., 2005; Stenevik and Sundby, 2007).

Filin and Oganin (2008) reported on a quantitative analysis of the response of the cod stock in the Barents Sea to future climate changes using a fish-ecosystem coupled model that takes into account trophic interactions and environmental influences. Their model included cod as a predator and only capelin and juvenile cod as a prey species. Results of the simulations indicated that raising the temperature in the Barents Sea by 1°-4°C would result in increased cod growth and earlier maturation. This would have a positive effect on the general production of the cod stock, but cannibalism would also increase.
While these possible changes for Atlantic cod under future warming are consistent with past observations, the actual response remains uncertain. Indeed, Drinkwater (2005) noted that improved understanding of the physiological and behavioural responses of cod to changes in environmental conditions, as well as the responses of other components of the marine ecosystem, are required for future cod projections. It is expected that cod growth rates will eventually decrease at very warm temperatures, and cod movements will be the result of complicated behavioural responses of individuals to multiple cues, gradients and variation in temperature, and fluctuations in stock abundance. Competitors and predators will also influence the response of the cod to climate change. For example, if Atlantic mackerel (Scomber scomberus) expand from the Norwegian Sea into the Barents Sea in substantial numbers, it could significantly reduce numbers of larval and juveniles cod through predation. It must be pointed out that the ultimate fate of cod will also depend upon fisheries and their management.  A possible build-up in cod biomass in the Barents Sea will dependent greatly on fishing intensity (Drinkwater, 2005, 2009).  Examining the effect of different management regimes on Norwegian cod fisheries in conjunction with climate change, Eide (2008) concluded that fishery management schemes will play a more significant role than climate change on future economic performance of the fishing industry in the Barents Sea.

Vikebø et al. (2007) examined the potential impact of a reduction in thermohaline circulation (THC) in the North Atlantic on the drift, growth, and distribution of larval and juvenile North-east Arctic cod. The THC brings warm water north which cools, sinks, and returns as a deep water current. Using the Regional Ocean Modelling System (ROMS), they imposed a 3-fold increase in river discharge to the Nordic Seas and the Arctic Ocean, which reduced the strength of the THC by 35%. This is near the projected reduction of around 25% in the THC predicted by the end of the 21st century in the IPCC (2007) report. Vikebø et al. (2007) found that this reduction results in fewer juvenile cod being transported eastwards in the Barents Sea, and those that were transported were considerably smaller in body size because of cooler temperatures. More juvenile cod were transported to western parts of Spitsbergen where feeding conditions are poorer and survival rates are lower than in the Barents Sea, which led to an overall increase in cod mortality for the year class. Some juveniles would possibly move into the Arctic Ocean, but were not expected to survive.
Although arguably cod is the most studied species in the Barents Sea in relation to the impacts of climate change, investigations of other species suggest that they will be substantially impacted as well.  Similar to cod, under projected warming in the Barents Sea, other boreal fish species are expected to extend farther east and north (Drinkwater, 2005; Loeng et al., 2005; Stenevik and Sundby, 2007; Hop and Gjøsæter, 2013). The eastward expansion of herring, blue whiting, and possibly Atlantic mackerel are expect to result in new species interactions and potentially to changes in the structure and function of the ecosystem. For example, substantial numbers of mackerel could reduce the cod population through predation on cod larvae and juveniles. Blue whiting has already increased significantly in the Barents Sea, but has not yet become a significant prey for piscivorous fish (Dolgov et al., 2010).  Capelin has been observed to have moved northward in response to the present warm temperatures and reduced ice (Ingvaldsen and Gjøsæter, 2013).  Cheung et al. (2009) suggested that a relatively large number of invasive species can be expected in the Sea Barents, as well as local extinctions and species turnovers related to distribution shifts. Polar cod is one species that has been predicted to disappear from the Barents Sea within approximately 30 years due to warming temperatures and reduction in sea ice (Cheung et al., 2008). Hop and Gjøsæter (2013) suggest that the polar cod may remain in the Barents Sea, but would lose the ice-associated part of its life cycle, and its summer distribution area would shrink significantly.  Climate change is also expected to result in higher overall production and subsequent increased catches of haddock, herring, and other boreal species (Loeng et al., 2005). Salmon abundance likely will increase in Russian waters as previously observed under warm conditions (Lajus et al., 2005), and also extend its range to northern Svalbard. Eriksen et al. (2012) has noted recent increase in jellyfish abundance in the Barents Sea in association with warmer waters, and more jellyfish are likely to appear as temperatures continue to increase. 
Possible impacts on the capelin population under climate change were first explored by Huse and Ellingsen (2008). They applied a combined physical and plankton model coupled with increased temperatures based on a previous IPCC B2 scenario. In addition to a projected distributional shift northwards during summer feeding in warm periods, new capelin spawning areas were also projected to be established off Novaya Zemlya and possibly off Eastern Svalbard.  Spawning would also occur earlier in the year. The potential of a shift in capelin spawning sites off Novaya Zemlya had been predicted earlier by Øiestad (1990).

Marine Mammals and Seabirds

Seal species which breed and raise their young on or near the ice edge — such as the ringed seal — would experience a loss of habitat under climate change. Durner et al. (2009) suggested that within the Arctic, the greatest rate of loss of polar bear habitat over the 21st Century would occur in the Barents Sea with a rate of 6.5% per decade.  Polar bears, which hunt seals near the ice edge, would have to move further north in search of prey. The earlier spring ice break-up and later fall freeze-up would force the polar bears off the ice earlier in the spring, or be left to deal with an unstable ice edge leading to a general reduction in body condition. Female bears would have to go longer distances in pursuit of food leaving cubs unattended and vulnerable (Stirling et al., 1999). Walruses and whales, which rely on sea ice of a relative thickness that they can break through to create breathing holes, would benefit from a thinner ice sheet, but walruses would then encounter the problem of finding adequate sea ice to support their body weight during resting periods.

The retreat of sea ice will threaten the existence of polynyas. These areas of high productivity are known for attracting large numbers of sea birds and marine mammals. In the Barents Sea, well known polynyas occur in Storfjorden and Hinlopen on the eastern part of Svalbard and southwest of Franz Josef Land. The loss of polynyas has traditionally been a result of the open water not appearing because of closure by surrounding ice. With climate change, however, the loss would most likely be caused by a lack of sea ice that helps to define polynyas. The importance of polynyas for biodiversity and productivity within the Barents Sea has not been quantified, thus how their disappearance will affect the region’s biodiversity is uncertain.
Species displacements may have negative impacts on seabirds and marine mammals which are used to feeding on specific prey. Fish and seabirds may alter their range in an attempt to locate suitable prey or adapt to a different food source. This could result in recruitment failure. Generally, seabirds feed only 100 km from their breeding sites but this range may be extended.

It must be cautioned that not only are the atmospheric and ocean climate scenarios highly uncertain, but their impacts are as well (IPCC, 2013; IPCC, 2014a, b; Howell et al., 2013).  Also, climate change is just one of the global change issues that marine environments are subjected to; other issues, such as fishing and ocean acidification, will also play roles and must be taken into account for projections of future conditions.  Several has projections were based on previous relationships between climate and food webs.  However, it has been noted by Bogstad et al. (2013) for example that we cannot count on such relationships continuing into the foreseeable future, as many of them are time variant.  With the high uncertainty and our lack of ability to forecast the future accurately, we must develop management strategies which are robust to unforeseeable changes in stock dynamics (Howell et al., 2013).