Climate in the Barents Region
Summary and Keywords
The Barents Sea is a region of the Arctic Ocean named after one of its first known explorers (1594–1597), Willem Barentsz from the Netherlands, although there are accounts of earlier explorations: the Norwegian seafarer Ottar rounded the northern tip of Europe and explored the Barents and White Seas between 870 and 890 ce, a journey followed by a number of Norsemen; Pomors hunted seals and walruses in the region; and Novgorodian merchants engaged in the fur trade. These seafarers were probably the first to accumulate knowledge about the nature of sea ice in the Barents region; however, scientific expeditions and the exploration of the climate of the region had to wait until the invention and employment of scientific instruments such as the thermometer and barometer. Most of the early exploration involved mapping the land and the sea ice and making geographical observations. There were also many unsuccessful attempts to use the Northeast Passage to reach the Bering Strait. The first scientific expeditions involved F. P. Litke (1821±1824), P. K. Pakhtusov (1834±1835), A. K. Tsivol’ka (1837±1839), and Henrik Mohn (1876–1878), who recorded oceanographic, ice, and meteorological conditions.
The scientific study of the Barents region and its climate has been spearheaded by a number of campaigns. There were four generations of the International Polar Year (IPY): 1882–1883, 1932–1933, 1957–1958, and 2007–2008. A British polar campaign was launched in July 1945 with Antarctic operations administered by the Colonial Office, renamed as the Falkland Islands Dependencies Survey (FIDS); it included a scientific bureau by 1950. It was rebranded as the British Antarctic Survey (BAS) in 1962 (British Antarctic Survey History leaflet). While BAS had its initial emphasis on the Antarctic, it has also been involved in science projects in the Barents region. The most dedicated mission to the Arctic and the Barents region has been the Arctic Monitoring and Assessment Programme (AMAP), which has commissioned a series of reports on the Arctic climate: the Arctic Climate Impact Assessment (ACIA) report, the Snow Water Ice and Permafrost in the Arctic (SWIPA) report, and the Adaptive Actions in a Changing Arctic (AACA) report.
The climate of the Barents Sea is strongly influenced by the warm waters from the Norwegian current bringing heat from the subtropical North Atlantic. The region is 10°C–15°C warmer than the average temperature on the same latitude, and a large part of the Barents Sea is open water even in winter. It is roughly bounded by the Svalbard archipelago, northern Fennoscandia, the Kanin Peninsula, Kolguyev Island, Novaya Zemlya, and Franz Josef Land, and is a shallow ocean basin which constrains physical processes such as currents and convection. To the west, the Greenland Sea forms a buffer region with some of the strongest temperature gradients on earth between Iceland and Greenland. The combination of a strong temperature gradient and westerlies influences air pressure, wind patterns, and storm tracks. The strong temperature contrast between sea ice and open water in the northern part sets the stage for polar lows, as well as heat and moisture exchange between ocean and atmosphere. Glaciers on the Arctic islands generate icebergs, which may drift in the Barents Sea subject to wind and ocean currents.
The land encircling the Barents Sea includes regions with permafrost and tundra. Precipitation comes mainly from synoptic storms and weather fronts; it falls as snow in the winter and rain in the summer. The land area is snow-covered in winter, and rivers in the region drain the rainwater and meltwater into the Barents Sea. Pronounced natural variations in the seasonal weather statistics can be linked to variations in the polar jet stream and Rossby waves, which result in a clustering of storm activity, blocking high-pressure systems. The Barents region is subject to rapid climate change due to a “polar amplification,” and observations from Svalbard suggest that the past warming trend ranks among the strongest recorded on earth. The regional change is reinforced by a number of feedback effects, such as receding sea-ice cover and influx of mild moist air from the south.
Climate can be defined as weather statistics that give a picture of the type of weather that can be expected. It can for all intents and purposes be quantified in terms of probability density functions (pdf) which describe the probability associated with the different weather states. This pdf can be regarded as inherent to the system and determined by the physical processes present. In order to estimate the parameters describing the pdf, such as the mean value, one needs a large statistical sample. Often this means statistics of weather over long periods of time, but it doesn’t always have to. The important point is that the statistical sample is sufficiently large for a good estimation of the statistical parameters of interest. The statistical properties (statistical distribution) of temperature have a bell shape, and can be described with the well-known normal distribution. The shape of the curve is mainly determined by two parameters: the mean and the standard deviation. Other weather elements have other statistical shapes, and the curve for precipitation is often based on the gamma distribution. Other statistical aspects may be typical time scales, spatial extent, and trends. One difference between weather and climate is that weather is unpredictable after some days, but weather statistics are often more predictable. For instance, the mean temperature varies systematically with latitude, altitude, and seasonality, whereas the mean precipitation is influenced by the distance to the coast and mountain ranges.
Traditional climate description based on Köppen−Geiger Climate Classification types places the Barents region in the categories polar tundra (ET) and snow-hot-dry summer (Dsa) for the the early 20th century. The region is characterized by a stark contrast between dark and cold polar nights and months with midnight sun. The temperature can exceed +30°C on calm and cloudless summer days when the heating from the sun does not cease with a sunset or as a result of warm air from the south being advected far north. During the dark and long winter nights the temperature may plunge to below −50°C in winter (Figure 1).
The Observational Network
Most of our knowledge about the climate in the Barents region comes from observations, monitoring, and measurements made with a range of scientific instruments. There are a number of station measurements, such as thermometer huts and rain gauges recording temperature and precipitation over time (Figure 2). Other types of station data include barometric pressure, snow depth, wind speed and direction, humidity, cloudiness and sunlight (shortwave radiation). Some of these records extend back in time over many decades, while some only cover recent years. These climatological data are supplemented by measurements carried out through expeditions and experiments as well as measurements made by instruments on board satellites (often referred to as remote sensing). In recent years, a number of boreholes have been drilled that can provide information about past temperatures which have caused heat to diffuse to greater depths in the soil.
Satellite data provide the best picture of the sea ice extent (Figure 3), but there are satellite-borne instruments which can measure sea surface temperature (SST), wind/waves (scatterometer), clouds, gases, and aerosol particles (pollution). The satellites can only measure different forms of light and rely on various algorithms (simple models) to estimate physical properties which have caused or interfered with electromagnetic radiation (e.g., scattering of light by particles).
Observations made on land are complemented by oceanographic observations made during scientific cruises, a fleet of autonomous floaters (ARGOS), and from fixed buoys. Some observations have also been made by conductivity–temperature–depth (CTD) gauges fixed to seals. Typical properties include temperature, salinity, and chemical tracers (e.g. CFC) and properties (pH).
Synthesis of Climate Information
The sparse population in the Arctic and the Barents region and the harsh environment make it expensive to maintain a dense network of measurements and explain the paucity of the data records there. The observational data can be combined with atmospheric models to fill in the gaps in a product known as an analysis or a reanalysis. The analysis is basically a weather forecast that has been initialized with the best available observations and models. A reanalysis is similar, but with a fixed model for the whole time span. One problem with reanalyses has been the paucity of data in the Arctic and the lack of constraints applied to the models in the simulation of the atmospheric conditions, known as “assimilation.” Hence, the most reliable data from the Arctic involve direct observations.
Present and Past Climate Conditions in the Barents Region
Annual mean surface temperature (at 2 m) averaged over 1961–2015 suggest that the annual mean temperature ranges from about +6°C along the coast of northern Norway to −18°C at Novaya Zemlya (Figure 4). There may be some slight variations in the estimates of the mean temperature in different reports, due to the use of different reference periods of data selection. The map of annual mean temperature shown in Figure 4 was a result of the temperature records in Figure 2 subject to a gridding procedure (kriging).
Most precipitation falls along the coast of northern Norway (~1,700 mm/year), but the annual amount is typically only 100 mm/year on Svalbard and Novaya Zemlya (Figure 5). There is more to precipitation than just the amount that falls over a period. Precipitation can be characterized in terms of six different aspects: its intensity (the wet-day mean precipitation, μ), its frequency (fw), how long the rains typically last (wet-spell duration, nnwd), typical dry-interval duration (number of consecutive dry days, ncdd), the spatial extent of the precipitation, and its phase (snow, sleet, rain, hail).
Mean Wind Conditions
The mean wind field based on the ERAINT reanalysis suggests changing directions of the prevailing wind from one region of the Arctic to the other. The wind direction is mainly southwesterly along the northern coast of Norway, but from the northeast over the northeastern part of the Barents Sea. The strongest wind speeds are found over the Greenland-Iceland-Norwegian (GIN) Sea (red area) and the open sea, with lower speeds over the sea ice–covered high north.
Geographical and Seasonal Climate Variations
The climate may be influenced by different factors, but some statistical properties of weather may be more sensitive to variations in conditions than others. The mean seasonal variation and the geographic differences in the seasonal cycle are due to physical conditions present, and the degree of similarity gives an indication of the sensitivity of the various statistics to changes in the physical environment. This sensitivity is visible in their response to different geographical conditions and the seasonal march (Figure 7). The seasonal variations in temperature range from −30°C to +20°C, with the greatest seasonal range in the interior of northern Russia (Figure 7a). There is also a clear influence of geographical factors visible, as the blue curves (the islands in the high Arctic) show systematically lower temperatures than the red ones (northern Fennoscandia). In other words, the mean temperature exhibits a systematic response to latitude, elevation, and the seasonal variations in the solar heating.
The daily temperature standard deviation also varies with location and season (Figure 7b), with the most pronounced variations in northern Russia and during winter (standard deviation σ ~ 10°C). The winter temperature is more variable than the summer temperature, and the standard deviation of the day-to-day temperature is greatest during winter and lowest in summer.
The persistence in temperature can be quantified by the day-to-day autocorrelation, which is close to 0.8 regardless of location and season (Figure 7c). All the curves tend to cluster and show the same flat structure, which indicates little sensitivity to variations in the solar angle or geographical factors. The day-to-day temperature persistence is therefore insensitive to both geographical variations and the seasonal variations.
In summary, both the mean temperature and the standard deviation exhibit sensitivity to the physical conditions, whereas the persistence is not sensitive to these factors. We can expect the two former to change over time if the physical environment changes, but the persistence is likely to stay the same.
Precipitation exhibits profoundly different properties from temperature, because there are dry days with no precipitation and wet days with variable amounts. It makes sense to sort the precipitation into dry and wet days, since different physical conditions are present for these two categories. The mean precipitation for wet days gives an indication of the precipitation intensity (Figure 7d), and for most locations the intensity is greatest during summer. Autumn storms are responsible for the most intense precipitation along the coast of Northern Norway, however. The frequency of wet days is strongly affected by the geography as well as the seasonal variations (Figure 7e). However, the character of the seasonal variation varies over the Barents region, and there seems to be less of a common pattern. The most common trait seems to be more days with precipitation in the cold season. There is also large variations in the typical duration of interval between each event with precipitation, although it is harder to discern a common trait for this than some of the other parameters. Springtime tends to have the longest intervals between wet days for most locations, but this too varies strongly from location to location (Figure 7f).
The response seen in various statistical parameters to changes various factors may indicate a high sensitivity in general, associated with changes in the energy balance. This sensitivity can be assessed through the analysis of historical trends and then compared to the sensitivity to seasonal variations and geographical factors.
Maps based on station observations (ECA&D) indicate substantial warming over the Barents region over the 1961–2015 interval, and the annual mean temperature increase has ranged between +1.4°C and +4.4°C (Figure 8). The annual mean temperature has increased by +1–2°C over the period 1954–2003 according to ACIA (2004), and the strongest warming has been observed in winter. The lower estimate for change in temperature over the period 1954–2003 compared to 1961–2015 suggests that the temperature change is not monotonous. The lower trend estimate for 1954–2003 is due to a warm period in the region during the 1930–1950s (the early 20th century warming) and the strong warming in recent years. The warming has been subject to natural variations in addition to global warming. For some locations there are long temperature records, such as Svalbard Airport, which indicate high temperatures in the 1930s–1940s–1950s, mainly during autumn and winter (Figure 9). The cause of this intermittent Arctic warm period is not well understood, although there are several plausible explanations.
The most pronounced warming has been observed at the higher latitudes, consistent with the long-predicted polar amplification. This amplification is a result of the Arctic temperatures being sensitive to various conditions which involve phenomena that amplify an original change—so-called “positive feedback.” Examples of feedback mechanisms include changes in snow and sea-ice cover which affect the albedo and the fraction of the sunlight absorbed in the ground and increased vapor pressure with higher temperatures. Other factors include cloudiness, storm tracks, planetary boundary layer height, and ocean currents, although their roles are often not fully understood yet.
According to an ACIA (2004) report, the amount of precipitation in the Arctic has increased by 8% over the 20th century. In order to understand whether the increased precipitation is due to more days with precipitation or greater amounts during days of precipitation, it is necessary to examine the wet-day frequency (includes snowfall) and the wet-day mean precipitation (also includes snow). Trend analysis for the wet-day mean precipitation and frequency suggests an unambiguous increase in intensity, but the frequency has increased over Fennoscandia and decreased over northern Russia. For the precipitation intensity (wet-day mean precipitation), the most pronounced increases have taken place at the lower latitudes in Fennoscandia and least changes in the highest latitudes over the 1961–2015 period (Figure 10a). The number of wet days has increased over most of Fennoscandia but decreased over Northern Russia and the high latitudes (Figure 10b). The precipitation occurrence is strongly affected by circulation patterns and storm tracks.
As the temperatures have increased, an increasingly higher proportion of the precipitation has fallen as rain as opposed to snow. Despite the increase in precipitation, the amount of snowfall has decreased in the warmest parts of the Barents region, but has increased in the coolest areas. The shift in the phase may have had an effect on the measurements, as the capture efficiency for liquid precipitation is higher than for frozen precipitation, especially in windy conditions.
Figure 11 shows histograms of trend analysis of different statistical parameters for temperature and precipitation in the Barents region. The clearest trend is seen in the increasing annual mean temperature (Figure 11a), decreasing temperature variability (Figure 11b), and precipitation intensity (Figure 11d). The increase in the mean temperature is consistent with the map in Figure 8, except for some stations with decreasing temperatures (negative trends). The two analyses use slightly different selection of the data, where the histograms include any station with more than 30 years of data, whereas the maps are based on the interval 1961–2015. Hence the differences are due to natural variations. There has been little long-term change in the temperature persistence, which is consistent with the robust nature of the day-one autocorrelation presented in Figure 5c. The statistics of day-to-day temperature persistence presented in Figure 11, the lag-one-day autocorrelation, seem to be robust to variations in physical factors such as trend as well as geographical differences and seasonal variations. The number of wet/snowy days has increased in some regions and decreased in others, giving rise to a histogram with two peaks (Figure 11e). This picture is consistent with the map in Figure 10b and the notion that the occurrence of rain and snowfall is connected to changes in the atmospheric circulation. The results also suggests little overall trend in the duration of dry intervals, the number of consecutive dry days.
Future Outlooks in the Barents Region
Projection of Common Elements
Downscaled global climate model simulations (GCMs) provide a future outlook for temperature in the Barents region with the greatest warming in winter, in the interior parts of Fennoscandia, and at higher latitudes such as the Svalbard archipelago (Figure 12a). The projected temperature increase between 2015 and 2080 ranges between +3°C and 10°C in winter for the RCP4.5 emission scenario. The degree of warming depends on the future greenhouse gas emissions and the severity of the retreat of sea ice.
Natural variations are a prominent feature of the Arctic climate, due to large temperature gradients, storm tracks, changes in ocean currents, and the shifting of sea ice. The exact timing of a substantial part of these natural fluctuations is unpredictable, since they involve strong nonlinear dynamics which have a chaotic character—minuscule differences in the description of an initial state lead to totally different outcomes, a phenomenon popularly called “the butterfly effect.” Climate models can nevertheless reproduce the statistical character of the natural variations, even if they fail to predict when they occur and how pronounced they are, and a larger number of parallel simulations with slight differences in model set-up or initial conditions (known as ensembles) tend to produce a range of results which tend to envelope the actual outcome. One example is presented in Figure 13, where the symbols represent the annual mean temperature recorded on Svalbard and the colored clouds different ensembles simulations. No single model simulation captures the observed temperature evolution accurately, but the ensemble as a whole envelopes the observations, indicating a reasonable match. Model validation, however, involves more sophisticated statistical analysis.
The yellow cloud in Figure 13 represents an emission scenario which is roughly on the middle in terms of the coupled model intercomparison project 5 (CMIP5, RCP4.5). The red cloud above represents higher emissions (RCP8.5), whereas the green cloud shows a future with low emissions (RCP2.6). The highest emissions may result in more than 10°C annual mean warming according to these results, whereas the lowest emissions indicate a modest future warming. Again, the mean temperature exhibits a sensitivity to variations in the environment, in this case the future emission scenario as opposed to the seasonal cycle or geographic factors. One single climate model indicated cooling for the future as a result of a collapse of the thermohaline circulation (RCP4.5 and 2.6).
The IPCC reports highlighted the presence of a greater spread in simulated outlook (greater uncertainty) for the zonal mean temperature at high latitudes compared to low latitudes. This result arises because both of a statistical artifact due to smaller samples and because the outlook for sea ice is sensitive to the model calculations. The rate of warming is also influenced by changes in the frequency and character of natural processes, such as synoptic storms and clouds. These phenomena are reproduced by present global climate models, but their character is not always simulated accurately. For instance, the estimation of clouds and storm tracks tend to differ somewhat from the observed ones.
Future outlooks for precipitation in the Barents region point to higher amounts and a change in the proportions falling as snow and rain. Part of the increase is expected from increased area with open water as the sea ice retreats. Estimates suggest an increase by 20% between 1980–2000 and 2100. Precipitation in a warmer Barents region will be more as rain and less as snow. Since monthly and annual mean precipitation are the product of the wet-day frequency and the wet-day mean precipitation, the change may be a result of more rainy days or more rain when it rains. Projections also indicate an increased risk for more frequent rain-on-snow (ROS) events in the future (Hansen et al., 2014).
The future state of the sea ice is a joker in terms of the future outlook for the climate in the Barents region. Sea-ice affects a large range of phenomena and conditions, such as air-sea exchange of heat and moisture, the boundary layer, the way winds affect oceans, the albedo, and clouds. Sea ice is sensitive to ocean currents, ocean mixing, salinity, winds, precipitation, and temperature and involves spatial scales too small to be resolved by global climate models. Nevertheless, most model projections indicate a retreat in sea ice, which is relevant for the winter climate of the Barents Sea. In summer, the Barents Sea is more or less open water, so the trends are most important in winter.
Synoptic storms are traveling low-pressure systems associated with high winds, cloudiness, and precipitation. The trajectory tracing their movement is often referred to as storm tracks, which tend to cluster over the mid-latitude oceans stretching from Newfoundland into the Barents Sea. Storms reaching the Barents region tend to arrive from the southwest and bring with the mild and moist air. They are generated by sharp temperature gradients and wind shear through baroclinic instabilities. The generation of such storms is often known as cyclogenesis, and one common region of cyclogenesis is the interface between frigid polar air and the more temperate mid-latitude air. This interface is associated with the upper air jet stream, which is highly nonlinear and difficult to predict over a long time. Its meandering character can be described in terms of Rossby waves that can be forecast a few days ahead. Changes in sea surface temperature, sea ice, and stratospheric conditions may affect the jet stream and the generation and propagation of synoptic storms.
The storms represent a hazard, causing high waves and storm surges, and can trigger avalanches. Some of these synoptic storms have caused heavy rain or been associated with ROS events on Svalbard (Hansen et al., 2014). Avalanches have been triggered by such storms, especially when they bring heavy snowfall and snowpack builds up on the leeward side of mountains. One such event took place in Longyearbyen on Svalbard on December 19, 2015 (AMAP, 2017a).
Synoptic storms are more frequent in winter and in the western part of the Barents Sea (Figure 14). The storm activity is also related to the variations in the North Atlantic Oscillation (NAO), with more storminess associated with a high NAO index. The connection between the storms and the NAO can partly be explained by more low-pressure disturbances in the vicinity of Iceland during an active storm phase, which then affects the observations of the barometric pressure and hence the NAO index.
The future outlook for storms in the Barents Sea is somewhat ambiguous, according to AMAP (2017a). The global climate models are able to simulate such storms, but the simulated storm tracks tend to follow a west-east direction that doesn’t extend sufficiently toward the north. The storms also tend to be too weak. Some climate model simulations indicate that storm activity is likely to decrease over the Barents Sea in the future, but other calculations making use of statistical modelling suggest an increase. Several studies suggest weaker fronts and reduced temperature gradients in a warmer world, but increased temperature and humidity (due to retreating sea-ice cover) can also be seen as more fuel for such storms.
Some model simulations project increased wind speeds over the northern part of the Barents Sea, in particular during winter and autumn. The increase is connected to the retreat of sea ice and can be explained as a result of changes in surface friction and the air-sea heat exchange. The same model results also indicate a slight decrease in wind speed at lower latitude, but little change during summer. This simulation provides only a plausible picture of the future, and other equally valid simulation may give different indications. One single simulation for a limited time slice does not provide an account of the possible range of outcomes connected to natural variability, for which the exact evolution cannot be predicted. Furthermore, the climate models may differ slightly in their design, which may result in different answers. The future wind speed is expected to be influenced by the presence of sea ice, and different climate models provide different accounts of the sea ice. In addition, global climate models have a tendency to simulate more westerly flow than is observed over northern Europe,1 and the storm tracks are not in the right position. Hence, the outlook for wind should still be regarded as highly uncertain.
Polar lows are maritime phenomena involving violent vortices with a diameter of ~100 km that last only a few hours and travel some distance. They are formed when cold air over ice-covered regions meets mild air over open ocean. There are on average 13 polar lows per year, and the activity is highest between October and April. The database on polar lows is scarce, making trend analysis difficult.
It is expected that the number of polar lows will decrease with future warmer conditions over the Barents Sea, as the sea surface and the mid-troposphere temperatures are expected to change the static stability in disfavor of polar low formation (AMAP, 2017a). Polar lows are also sensitive to changes in the planetary boundary layer, sea ice, and snow.
There are a number of cloud categories associated with different phenomena, such as clouds associated with synoptic storms, convection, weather fronts, and fog. Observations of clouds in the Barents region suggest an increased frequency in the convective clouds, but a lack of significant change in the occurrence of shallow stratified clouds (AMAP, 2017a). Clouds are influenced by the planetary boundary layer, which is the layer of air above land and ocean through which the atmosphere and the medium below exchange moisture and energy. This air layer is often turbulent and affects some types of clouds (shallow convection, fog), sea ice, and temperatures (e.g., through inversion). Clouds are also associated with synoptic storms and are a result of mechanisms fueling the storms through latent heat release. They form as vapor condenses into cloud drops or ice crystals.
Clouds play a complicated role in the Arctic climate, where they act as an agent that soaks up the atmospheric vapor and precipitates it as snow or rain. Through the phase transition of water, they are part of the chain in which heat and energy flows. The clouds influence the planetary albedo, and can mask the albedo of underlying surfaces with snow or ice. They also affect the outgoing longwave radiation, and can act as an isolating layer during the polar night.
According to the IPCC, some of the greatest uncertainties associated with climate simulations involve clouds. They represent physical processes on a large range of scales, from cloud microphysics involving micrometer scales to 100 km cloud structures. They may contain a range of different water-based aerosols, such as droplets, snow, and ice crystals, which are sensitive to local temperature, moisture, and turbulence in addition to electric forces.
The stratosphere over the Arctic may have an impact on the climate in the Barents regions, and both analyses based on observations and model simulation indicate the potential stratospheric precursors which may influence weather statistics closer to the surface. One example is sudden stratospheric warmings, which can affect the atmospheric circulation. The stratospheric ozone affects the amount of UV radiation reaching the surface as well as temperatures above the tropopause. The stratospheric temperatures, winds, and densities may affect vertical wave propagation and have been hypothesized to be useful for seasonal forecasting.
Barents Sea Oceanography
The Barents Sea covers about 10% of the Arctic Ocean and has a mean depth of 230 m. It is the site of an intense heat and moisture exchange between ocean and the atmosphere as well as variations in sea ice, and a place where dense water forms which later sinks to the deep abyss of the Arctic Ocean. The deep, dense water is part of the thermohaline circulation and the ocean “conveyor belt” and involves the upper layer of ocean water subject to wind-generated mixing. The surface water plays a key role in the way the ocean is coupled to the atmosphere, and much of the air-sea exchange in the Barents Sea is connected to the Atlantic inflow of relatively warm water near the surface. The ocean surface loses substantial amount of heat to the atmosphere before it cools down. Salinity influences the water density, and as the water cools, the combined effect from temperature and salinity plays a key role in the deep water production through convection. The exchange between the atmosphere and ocean is also influenced by sea ice and lighter fresh water that can form a “lens” on top of more saline water from the Atlantic. Fresh water can come from ice melt, rivers, or precipitation. There is a Polar front in the upper ocean layers separating cold Arctic from surface currents of warmer Atlantic water.
The Barents Sea is an important region for the ecosystem, being home to more than 200 species of fish and a spawning ground for cod. It is also the richest region in terms of the population of marine mammals. The ecosystem is flourishing due to its oceanographic character, with inflow from the Atlantic and the interface between warm Atlantic and cold polar water. The sea ice also provides living conditions for a number of species, and the Barents Sea harbors one of the largest concentrations of seabirds in the world (AMAP, 2017a). The biological activity, however, is strongly tied to the seasonal cycle.
Changes in the Barents Sea
During summer, there is little sea ice in the Barents Sea, but part of the ocean is covered in winter. There are pronounced variations in the Barents Sea sea-ice cover from one year to the next. Nevertheless, there has been approximately a 7% decrease in the sea-ice cover over the 1979–2007 period. The reduction in sea ice influences the air-sea exchange of heat and moisture, as well as clouds, precipitation, and the wind fetch (waves). Wintertime reduction of sea ice is of particular importance, because it strongly increases the energy transfer from the ocean to the air, which will lead to weaker winter temperature inversion. Summertime reductions in sea ice over the Arctic affect the planetary albedo (reflection of sunlight) and the capacity of the ocean to absorb solar heating. Sea ice also influences the friction between the surface and the air, and hence the wind pattern. Some studies have indicated that a reduction of sea ice in the Kara Sea may give rise to anomalies in the atmospheric circulations and low and high pressure systems which may bring cold air over northern Europe. Similar mechanisms may potentially be present for the sea ice in the Barents Sea. Furthermore, model simulations focusing on seasonal predictability suggest that sea-ice anomalies may affect the subsequent weather conditions. Sea ice also stops sunlight from penetrating the upper ocean, and hence affects ocean temperatures as well as biosynthesis.
The global sea level increases as a result of global warming, where warmer water expands and takes up a larger volume, in addition to ice on land melting and adding water to the ocean. Locally, the sea level responds to the global change but is also influenced by winds, ocean currents, and gravitational forces. Ice melting on Greenland will add water to the world oceans, but its weight will diminish and the ground will rebound to a new static balance while the gravitational forces from the ice mass will be redistributed as the water flows into the oceans. For that reason, ice melt on Antarctica is expected to have a greater effect on the local sea level in the Barents Sea than equal ice melt on Greenland. The relative sea-level rise varies from place to place due to tectonic uplift (rebound from the past ice age), and varies from 2.5 to 4.0 mm/year along the coast of Northern Norway.
The sea level may rise abruptly if land-based or grounded ice sheets collapse and disintegrate into the oceans. Such an event will be a tipping point, as the process cannot easily be reversed. Increased sea level combined with storm surges also increases coastal erosion, a process which is not readily reversed.
Estimates of past and future changes in ocean wave heights are ambivalent, and different analyses point in different directions. Reduced sea-ice cover is expected to give longer fetch, but there is little indication of changes in wave height over the Barents Sea. There are also few indications of changes in wind speed over the past decades. However, an increasingly thicker layer of surface water is consistent with more waves and more stirring of the waters. The salinity near the ocean surface has been influenced by mixing in salty water from greater depths, compensated by fresher water from ice melt and precipitation. The rate of mixing is sensitive not only to wind speed but also to the vertical stratification of the oceans and the heat exchange with the atmosphere, as well as temperature and salinity. Another effect is a weakening of the vertical stratification of the water column, which can favor more localized winter convection and fewer large-scale overturning events (AMAP, 2017a).
Icebergs are different from sea ice; they are a product of calving coastal glaciers, whereas sea ice forms on the ocean surface when the water gets sufficiently cold and freezes. Icebergs are more numerous during times of warming after a cold period and when coastal glaciers melt and disintegrate. The sources of icebergs in the Barents Sea are mainly Novaya Zemlya, Franz Josef Land, and Svalbard. Iceberg drift is subject to winds and currents, and Russian scientists have observed an increasingly southward reach over a period of 57 years. Some explanations may be changes in the wind and ocean currents, but a higher number of icebergs also increases the odds for more widespread geographical sightings. The number of icebergs fluctuates from year to year, and the record is incomplete. A Russian record suggests a low number (less than 50 per year) before 1950 and larger variability since then (200–1,400 per year).
The ocean surface is supersaturated in terms of calcium carbonate at the present, but the state of saturation diminishes with more CO2 being dissolved. The degree of saturation affects marine creatures forming shells based on calcium carbonate. The ocean uptake of CO2 is influenced by sea-ice cover, biological activity, surface water temperature, water mixing, and ocean currents. At the observing site Station Mike (66°N/2°E), the recorded pH (which has a logarithmic scale) has dropped by 0.001 between 2001 and 2005. Cold water more efficiently absorbs CO2 than warmer water (Henry’s law), and the ocean uptake is higher at high latitudes with colder waters. The fraction of CO2 captured by the oceans reduces the additional greenhouse effect it would have if it were in the atmosphere. When the oceans get warmer, however, they may start releasing CO2 and become a source rather for a sink in terms of atmospheric CO2.
There are several aspects to snow, such as spatial extent (area), seasonal duration, depth, and quality (snow/ice/graininess). Snow accumulates in winter over land and on sea ice, and affects the exchange of heat and moisture between surface and the air in addition to the local albedo. It is also important for the hydrology, storing the water in a frozen state during winter and releasing it in spring. The geographical distribution and snow depth depend on snowfall, wind, and temperature. The snow quality is influenced by temperature and ROS events, and melt-freeze events create crusts and ice layers which are resistant to wind forces. Snow also influences vegetation and the ecosystem in the Barents region, and may become a problem for grazing animals if it has layers of ice within its structure. Snow can also be a hazard, and large accumulation in steep terrain can lead to avalanches.
Figure 15 shows a map of the typical annual maximum snow depth, revealing geographical variations. More snow is accumulated in the mountainous regions compared to lower elevation in Scandinavia, due to lower temperatures at higher elevations and more precipitation as a result of cooling of moist air forced aloft over higher ground (orographically forced precipitation).
The Barents region contains glaciers, which are land-based masses of ice that have resulted from accumulated snowfall and are offset by summer melt and calving. The ice forms when the snow is compressed by the weight of snow falling on top, or by freezing water. The glaciers bear witness to climatic conditions such as wintertime precipitation and temperatures as well as summertime temperatures. They occupy about 4% of the land area, mainly in Novaya Zemlya, Svalbard, and Franz Josef Land. Glaciers are part of the cryosphere as well as the hydrology, where seasonal melting impacts river runoff or the salinity of the coastal waters. Over geological time scales, land ice has been a factor in shaping the terrain.
The major rivers in the Barents region drain into the Barents Sea, adding fresh water into the ocean and reducing the salinity. The largest river by volume is the Pechora River west of the Ural Mountains, although it represents a smaller watershed (catchment) than the North Dvina River. Several rivers, especially in Northern Norway, have been dammed and regulated, and have an unnatural seasonal runoff.
Observations suggest a change in the seasonal character of the river runoff and the forming/disintegration of ice on the water waves. These changes can be connected to temperature and snowmelt. Winter and spring river discharge has been recorded in northern Finland, but summer discharge has decreased further south.
Permafrost is frozen soil and is widespread in the Barents region, but extends into variable depths depending on the ground conditions and the local climate. The ground is thawing in parts with warming where the temperature has exceeded freezing point. Thawing of permafrost is associated with methane release and has been observed on Svalbard and in northern Russia and northern Fennoscandia (AMAP, 2017b). Permafrost and subsurface temperatures can be monitored through boreholes. Variations in temperature near the surface propagate downward through a diffusive effect, and the shape of the temperature curve with depth can reveal past trends.
The vegetation is sensitive to climatic aspects such as the mean temperature, precipitation, the ranges between minima and maxima wind, and the frequency of weather states. About 20% of the Barents region consists of tundra, barren plains with no forest. The Arctic forest, often referred to as taiga, covers about 54% of the land area. In the west, the vegetation is Scandinavian Montane birch forest and grassland, although the Kola Peninsula is also covered by tundra. The vegetation influences the local climate, for example through albedo effects, surface friction, and evapotranspiration, in addition to itself being sensitive to the climate.
One definition of a tipping point is that changes in one direction happen over a short period of time whereas a reverse process takes much longer time (for all intents and purposes, an irreversible process with pronounced consequences). Examples are the disintegration of ice sheets or glaciers. The ice takes long time to accumulate, but can slide rapidly into the oceans. Other tipping points include the release of methane through the thawing of permafrost. There have also been some reports of methane leaking from gas hydrates at the Barents Sea ocean floor. Other types of tipping points may involve a thermohaline circulation collapse due to increased freshness of the Arctic Sea. There may be tipping points in the ecosystem where species become extinct, which can alter vegetation and subsequently the local climate.
I am grateful for the comments from Øyvind Nordli on this article.
AMAP. (2017a). Adaptation actions for a changing Arctic: Perspectives from the Barents area. Oslo, Norway: Arctic Monitoring and Assessment Programme (AMAP).Find this resource:
AMAP. (2017b). Snow, water, ice and permafrost in the Arctic (SWIPA) 2017. Oslo, Norway: Arctic Monitoring and Assessment Programme (AMAP).Find this resource:
Barr, S., & Lüdecke, C. (Eds.). (2010). The history of the International Polar Years (IPYs). Heidelberg: Springer.Find this resource:
Benestad, R., Parding, K., Isaksen, K., & Mezghani, A. (2016). Climate change and projections for the Barents region: What is expected to change and what will stay the same? Environmental Research Letters, 11(5),102170.R2.Find this resource:
Borodachev, V. Y., & Alexandrov, V. Y. (2011). History of the Northern Sea Route. In O. M. Johannessen, V. Y. Alexandrov, I. Y. Frolov, S. Sandven, L. H. Pettersson, L. P. Bobylev, . . . N. G. Babich (Eds.), Remote sensing of sea ice in the Northern Sea Route: Studies and applications (pp. 1–23). Berlin: Springer.Find this resource:
Dicks, L., Almond, R., McIvor, A., & Arctic Monitoring and Assessment Programme. (2013). Arctic climate issues 2011: Changes in Arctic snow, water, ice and permafrost. Oslo: Arctic Monitoring and Assessment Programme.Find this resource:
Hansen, B. B., Isaksen, K., Benestad, R. E., Kohler, J., Pedersen, Å. Ø., Loe, L. E., . . . Varpe, Ø. (2014). Warmer and wetter winters: Characteristics and implications of an extreme weather event in the high Arctic. Environmental Research Letters, 9(11), 114021.Find this resource:
Hassol, S. J. (2005). Impacts of a warming Arctic: Arctic climate impact assessment. Cambridge, UK: Cambridge University Press.Find this resource:
Klein Tank, A. M. G., Wijngaard, J. B., Können, G. P., Böhm, R., Demarée, G., Gocheva, A., . . . Petrovic, P. (2002). Daily dataset of 20th-century surface air temperature and precipitation series for the European Climate Assessment. International Journal of Climatology, 22, 1441–1453.Find this resource:
Nordli, Ø. (2010). The Svalbard Airport temperature series. Bulletin of Geography. Physical Geography Series, 3, 5–25.Find this resource:
Nordli, Ø., Przybylak, R., Ogilvie, A. E., & Isaksen, K. (2014). Long-term temperature trends and variability on Spitsbergen: The extended Svalbard Airport temperature series, 1989–2012. Polar Research, 33(1), 21349.Find this resource: