Longterm Changes In Some Water Mass Characteristics Of The Arctic Ocean

The current state of Arctic Ocean water and its long-term variability are important factors because they reflect changes in global climate and influence these changes. Salinity is of special significance because water density and hence water dynamics (currents, convection) depend on it at high latitudes. The vertical distribution of water density influences heat exchange between the ocean and the atmosphere, and this heat exchange plays a major role in the formation of atmospheric conditions, including atmospheric circulation and its cyclonic or anticyclonic nature. Salinity is also important as an indicator of the freshwater budget of adjacent areas. Hence, there is considerable interest in investigating the salinity of the Arctic Ocean and spatial-temporal patterns in its variability, especially in the surface layer where this variability is most strongly manifested.

Alekseev et al. (2000) calculated freshwater content relative to a referenced salinity of 34.8V (Aagaard and Carmack, 1989) and investigated the distribution of salinity in the upper 400-m layer of the Arctic Basin. The multiyear average freshwater content is maximum at the center of the anticyclonic Beaufort Gyre (approximately at 76°20'N, 152°W), where the freshwater content reaches 19 m, gradually decreasing to 1-2 m with increasing distance from the center of the gyre to its periphery. The origin of the Beaufort Gyre freshwater reservoir and its possible influence on Arctic Ocean circulation and climate are described by Proshutinsky et al. (2002), who show that the major cause of the large freshwater content in the Beaufort Gyre results from the process of Ekman pumping associated with climatological anticyclonic atmospheric circulation over the Canada Basin, centered in the Beaufort Gyre region.

Data available from many years of oceanographic observations, including broad surveys and data from drifting and polar stations, allowed us to calculate the mean annual freshwater content in the surface layer (up to 50 m), and obtain values of its multiyear changes within most of the Arctic Ocean. Because the amount of data available per unit area varied greatly in time and space, a technique for reconstruction of oceanographic characteristics was developed based on the spectral expansion method (Koltyshev and Timokhov, 1997). Using the oceanographic database of the Arctic Ocean, the temperature and salinity fields were reconstructed for March-May and a continuous series of characteristics was obtained for 1950 to 1993 at grid points with a spatial step of 200 km at standard oceanographic levels. The gridded fields of the reconstructed salinity values were used to calculate average salinity in the 5-50 m layer at the grid points, which made it possible to track changes in salinity for the indicated series of years (Ivanov et al., 2003; Gudkovich et al., 2004). According to these data, the freshwater content in the surface layer of the basin has decreased by approximately one-third over the 43-year time period. Surface water salinity is known to be influenced by many factors:

— Freshening due to river runoff.

— The balance of atmospheric precipitation and evaporation at the ocean surface.

— The balance of the processes of ice growth and melting.

— Processes of upwelling and downwelling near the shores and landfast ice under the influence of winds.

— Processes of convection and vertical turbulent mixing of waters.

— Salt advection by currents of different origins.

— The influence on the Ekman pumping layer of non-uniform wind fields (baric fields—cyclones and anticyclones) accompanied by upwelling (in cyclonic systems) and downwelling (in anticyclonic systems).

Changes in the intensity and direction of these processes in time and space results in corresponding salinity changes. The large-scale processes appear to be of greatest interest for studies of climate change because they have major long-term consequences. The last three processes listed are the most important, especially the last, while the first four either have a lesser broad-scale influence or are local in character.

Figure 4.19a, b shows changes in the distribution of average salinity in the 5-50 m water layer in the Arctic Basin, described by linear trends for the periods 1950-1988 and 1989-1993 (Gudkovich et al., 2004). At first glance, the distribution of salinity changes in space in the Arctic Ocean has a mixed character. However, some important repeated features are evident in these changes. During the time interval 1950 to 1988:

— A salinity increase was noted over much of the Arctic Basin, especially in the area of the Beaufort Gyre.

— The changes had the opposite sign (freshening) at the periphery of the gyre, in a zone extending from the north coast of Greenland to the New Siberian Islands.

— A salinity increase is also observed in the northern Greenland Sea and in the area adjoining the Barents Sea to the north, in the Pechora Sea, and in the northern Chukchi Sea.

— Freshening is also observed in the western Greenland Sea and in the Norwegian, Barents, Kara, Laptev, and East Siberian Seas.

A salinity decrease in the Arctic Seas (except for the Chukchi Sea) can at least be partly connected with increased runoff from large Asian rivers flowing to the Arctic Seas (the increase comprised approximately 10%). A salinity increase in the Chukchi and Beaufort Seas may result from corresponding changes in advection of Pacific Ocean waters. An increase in salinity over much of the Arctic Basin, especially in the Beaufort Gyre, is attributed to changes in atmospheric circulation: attenuation of the Arctic High and increased recurrence of cyclonic fields (see Section 4.2).

We remind readers that the upwelling that results in increased surface-water salinity in the central areas of baric depressions is accompanied by downwelling and freshening at their periphery. This probably explains the presence of a freshening belt extending from Greenland to the New Siberian Islands and the Siberian shelf seas (Figure 4.19a). Salinity change in the Greenland Sea is similar: deepening of the Iceland trough leads to salinification of surface water in the area known as the cold water dome, located in the northern part of the sea, and to freshening at the periphery of the cyclonic gyre in the Norwegian Sea (Gudkovich and Kovalev, 2002a). Intense freshening during the second half of the twentieth century was confirmed by Belkin et al. (1998) using direct-observation data taken onboard the weather ship (66°N, 2°E) in the southeastern Norwegian Sea.

As Figure 4.19b shows, during the next relatively short time interval (1989-1993), there were significant changes in the distribution of salinity trends in the surface layer of the Arctic Basin. A zone of increasing salinity moved westward, overlapping the area where freshening was previously observed. Salinity began to decrease in the area adjoining the north shores of the Canadian Arctic archipelago. A zone of freshening also appeared to the north of the Barents Sea. The main cause of these changes was significant modification of the pressure field expressed in a substantial decrease in atmospheric pressure in the region between Greenland and the Laptev Sea (Figure 4.20). As might be expected, salinification of surface waters in this region due to upwelling was accompanied by freshening at its periphery.

Figure 4.19. Changes in the distribution of

Figure 4.19. Changes in the distribution of average salinity in the 5-50-m water layer described by the linear trends for the periods 19501988 (a) and 19891993 (b). (1) 100-m isobath. (2) Salinity increase. (3) Salinity decrease.

Figure 4.20. Average difference in atmospheric pressure (hPa) between the periods 1985-1995 and 1970-1980 for January-March (a) and July-September (b).

35 20

1960 1970





35 20




1960 1970

1960 1990


1960 1970

1980 1990

Figure 4.21. Change in average salinity in the 0-100-m layer in the Norwegian Sea from weather ship M observations (a) and in the 0-50-m layer in the Arctic Basin (b).

As described above, the multiyear series of salinity values in the surface water layer (5-50 m) of the Arctic Basin in the March-May period at regular grid points allowed us to calculate average values (Savg) for the entire basin. Figure 4.21b plots changes in the Savg value from 1950 to 1990, with the actual fluctuations approximated by a polynomial to the power of 6; there is a pronounced linear trend pointing to a gradual salinity increase in the basin from the beginning to the end of the time interval. Cyclic changes are also identified:

— Increase in salinity in the late 1950s-early 1960s.

— Subsequent decrease in salinity in the late 1960s-early 1970s.

— A new increase in salinity around 1980, which was replaced by its decrease at the end of the time series.

The period of observed cyclicity lasts about 20 years. There are also higher frequency fluctuations with periods of 6-10 years.

It is reasonable to compare the salinity changes in the surface water layer of the Arctic Basin with the corresponding processes in the Norwegian and Greenland Seas, where salinity fluctuations of the same frequency were revealed and explained by the presence of self-oscillations in the ocean-ice cover-atmosphere system (Gudkovich and Kovalev, 2002a) (also see Section 5.3).

Figure 4.21a plots salinity changes within the 100 m horizon for the Norwegian Sea from weather ship M observations. Comparison of Figures 4.21a and b indicates a surprising interrelationship between the two areas. Because observations onboard the weather ship in the Norwegian Sea were made at the periphery of the cyclonic gyre, where the water density changes are opposite in sign to the changes occurring in the central zone of the gyre (Gudkovich and Kovalev, 2002), it can be concluded that long-term salinity changes in both regions had a similar character. Salinity fluctuations in the surface layer of the Arctic Basin exhibit a 3.25-year lag, on average, compared to the changes at the center of the cyclonic gyre of the Greenland Sea (Table 4.7).

Table 4.7. Years of salinity maximums and minimums in the upper water layer, vorticity of the wind fields, and corresponding lag values in two regions of the Arctic Ocean, the Greenland Sea (GS), and the Arctic Basin (B).


Salinity (S)

Vorticity of the wind fields (AP)

T, years

(AP - S)ab



t, years (AB - GS)



T, years (AB - GS)















































It is useful to estimate the time lag of average salinity variations in the surface layer of the Arctic Basin relative to the baric field changes over it. As an indicator of the latter, the large-scale vorticity index values were calculated using the Laplacian of mean atmospheric pressure at ^ = 84°N, A = 130.2°E:

where Pt, P0 are atmospheric pressure at the points located at the square angles and at its center, respectively for Region 0 (see Figure 4.22).

Figure 4.23 shows a change in time for the average values of this index for March-July smoothed by running 11-year periods. There is a clear 20-year cycle in the salinity variations in this index. Fluctuations in this index are generally close in phase to the changes in the high-latitude zonality index (Iz) considered above (Figure 4.7): the location of the extremes differs within ±2 years.

North Atlantic cyclones are known to become significantly deeper in the region of the Norwegian Energy Active Zone (NEAZO), which is connected to an important center of atmospheric circulation, the Icelandic depression. Hence, cyclones spread to the northeast and east, influencing the formation of the baric field over the Arctic, northern Europe, and Asia.

When the cyclones move to the northeast (Vangengeim-Girs circulation—type E), a baric trough formed over the Arctic Basin and the seas from the Barents to the Laptev becomes deeper with increased cyclonic activity in NEAZO, and is partly filled upon weakening. By slightly simplifying the processes, it can be concluded that vorticity of the wind fields in the area of the Icelandic depression and in the Arctic should change quasi-synchronously or with a small lag along cyclone pathways.

60 90 110

60 90 110

60 90 120

Figure 4.22. Layout of the regions for calculation of vorticity index J0, J1, and J2 values.

60 90 120

Figure 4.22. Layout of the regions for calculation of vorticity index J0, J1, and J2 values.

When cyclones move to the east (circulation type W), the baric trough is mainly formed over northern Europe and northern Asia, and the Arctic High should intensify. Hence, changes in wind-field vorticity and corresponding atmospheric pressure anomalies in the Arctic and in the region of the Icelandic depression will mainly occur in opposite phase. Under real extended conditions, there is usually cyclone movement both to the east and northeast (there are only changes in recurrence of different trajectories). Thus, some phenomena occur during changes in wind-field vorticity and corresponding atmospheric pressure anomalies, when processes in the Arctic Basin precede changes in the Iceland depression area. Exactly such a situation is apparent in a comparison of maxima and minima in 20-year cyclic changes in baric fields (Table 4.7).

Table 4.7 shows that the maxima and minima of cyclonic activity in the Arctic Basin precede (on average by 5 years) similar maxima and minima in the Greenland Sea. If the salinity response in the Greenland Sea to atmospheric conditions has an average lag of 2 years, the same response in the Arctic Basin takes about 10 years. The

AP, hPa 4t

Figure 4.23.

Variation over time of the vorticity index value of J0 — AP at the point of p — 84°N, A — 130.2°E for March July using

Figure 4.23.

Variation over time of the vorticity index value of J0 — AP at the point of p — 84°N, A — 130.2°E for March July using

1935 1945 1955 1965 1975 1985 1995 11-year running averaging.

differences must be caused by differences in the intensity of atmospheric processes in these regions. This result confirms the earlier conclusion by Gudkovich and Nikiforov (1965) regarding the significant stability of large-scale water circulation in the Arctic Basin.

The results of the present study show that 20-year cycles of change in the Greenland Sea generate corresponding changes in the Arctic and probably in the North Atlantic where a 21-year cycle is evident in the spectral density function of the NAO index (Figure 4.11). The trends in multiyear large-scale processes in the atmosphere and the ocean provide evidence of the influence of natural and possibly of anthropogenic factors or of cyclic fluctuations lasting longer than 100 years. Both factors may be at work here, along with processes taking place in the energy-active zone of the Greenland Sea.

Swift et al. (2005) studied the variability of different Arctic Basin water masses during the second half of the twentieth century by subdividing the basin into 13 boxes and averaging the water properties in each box. Based on their findings, these authors posed alternative hypotheses to the explanation of the main cause of surface water salinification.

As noted above, surface water salinity in some seas depends on their sea ice extent. The inverse character of the relationship between sea ice extent and salinity was confirmed by observations in the Kara and Chukchi Seas (Gudkovich et al., 1972, 1997). This relationship was observed in spite of the fact that the water-ice phase transitions should have resulted in the opposite changes: increased growth and decreased ice melting in the "cold" epochs should have resulted in salinification, and the decreased growth and increased melting in the "warm" epochs should have had the opposite effect. As shown in Appel and Gudkovich (1984), salt advection by ocean currents has a much greater influence on salinity (for example, the flow of relatively saline Barents Sea water to the Kara Sea through Makarov Strait and transport of Pacific Ocean water through Bering Strait to the southwestern Chukchi Sea with the Long Strait branch of the Bering Sea current).

These patterns disprove the widespread opinion expressed in scientific publications that the Arctic warming that began at the end of the twentieth century is accompanied by freshening of Arctic Ocean surface water, increased outflow of Arctic Ocean surface water to the North Atlantic, and a corresponding influence of Arctic Ocean surface water on thermohaline circulation in this region (e.g. Hassol, 2004, etc.). As demonstrated above, due to a weakened Arctic High (increased cyclonic activity in the Arctic) and the inflow of more saline oceanic waters from the south, the salinity of surface water over much of the Arctic Basin during this period has increased. Ice export from this basin to the Greenland Sea has decreased (see Section 4.4.1). This could lead not to freshening but rather to salinification of North Atlantic water. We think that the area of decreasing salinity in the North Atlantic has been limited to the northwestern region of the Atlantic Ocean adjoining Davis Strait, located behind the Icelandic depression.

The peak for river runoff and ice formation and melting processes in the Norwegian Sea (minimum salinity percentage) is known to have occurred at the end of the 1970s (see Figure 4.21a). At the end of the period of Arctic cooling, ice export from the Arctic Basin to the Greenland Sea intensified, resulting in freshening of the water in the northeast Atlantic and in the North European Basin. Thus, descriptions of the processes related to climate warming at the end of the 20th century were significantly misinterpreted by those who claim that this natural phenomenon will prove to be catastrophic if left unchecked.

Water temperature is also quite significant in the processes of climate change. Temperature profiles of different Arctic Basin water masses are presented in studies by Timofeyev (1960), Frolov et al. (2005), Polyakov et al. (2004, 2005), and others. In the surface layer of the Arctic Basin, the water temperature is close to the freezing point of water of relevant salinity. However, in the deep layers that contain relatively warm water of Atlantic origin, the temperature depends on both the volume and temperature of incoming Atlantic water. The effects of climatic change on these parameters were observed by Zubov (1938), who noted that during the period of Arctic warming in the 1920s-1930s, the average water temperature of the Nordkapp current (0-200 m layer) was 0.7-0.8°C higher than at the beginning of the twentieth century. Based on observations of the 1937-1940 expedition aboard the icebreaker G. Sedov, Timofeyev (1960) showed that the average temperature in the Atlantic water layer of the Arctic Basin was much higher than that measured in the same region by the 1893-1896 Fram expedition, while the maximum temperature of this layer increased by 0.7-1.0°C. There were significant interannual fluctuations in the mid-twentieth century as the temperature began to decrease. This was confirmed by Bulatov and Zakharov (1978), who investigated changes in the thermal state of the Arctic Ocean for a 20-year period from the mid 1950s to the mid 1970s. These authors observed some cooling of Arctic Basin waters, consistent with atmospheric cooling during this period. This cooling was more pronounced in the sub-Atlantic sector of the basin than in its sub-Pacific Ocean sector. Lamb and Johnson (1964) found that the temperature of deep Atlantic water was even lower at the peak of the Little Ice

Age (1790-1829), when the water temperature at the surface of the North Atlantic was 2-3°C lower than the current temperature.

Alekseev et al. (1998) focused their study on a significant new increase (anomaly of about +1°C) in the deep Atlantic water temperature in the 1990s that had been previously revealed by Quadfasel (1991) and Schauer et al. (1995). The upper boundary of this water mass was significantly higher, and the layer was thicker. An analysis of these observations suggests that each twentieth-century cycle of climate warming was accompanied by an increase in the temperature and the heat content of the deep Atlantic layer in the Arctic Basin. The quantitative indicators of warming of this water provide evidence that the 1990s warming event was similar to the 1930s Arctic warming. Note that comparing the anomalies of Atlantic water temperature in different years requires accounting for the location of the main flow of this water near the Eurasian continental slope, where the anomalies of climatic changes are maximized; they decrease rapidly toward the Canadian Arctic archipelago (Alekseev et al., 1998).

Alekseev et al. (1998) identified a significant phenomenon that accompanies warming: an increase in salinity of the upper water layer. This results in a decrease in the vertical density gradient and a corresponding increase in the heat flux from depth, which can contribute to a decrease in sea ice thickness along with an increase in air temperature and snow cover thickness during epochs of climate warming. As noted in Section 3, the intensity of ice growth observed onboard G. Sedov was 20% less than that observed during the Fram drift.

In an interesting study, Polyakov et al. (2004) analyzed a large set of observational data on the changes in Atlantic water temperature in the Arctic Basin during the twentieth century and found that low-frequency fluctuations with a period of about 60 years are clearly evident (Figure 4.24). Coherence is shown in the fluctuations of Atlantic water temperature and surface air temperature, ice cover in the Arctic Seas, and thickness of landfast ice in the vicinity of polar stations. The increased inflow of warmer Atlantic water through the Norwegian Sea is accompanied by changes in water density distribution in the Arctic Basin, which, along with the increased outflow of cold and freshened water to the North Atlantic through Davis Strait, reduces low-frequency fluctuations in the atmosphere-ocean-ice cover system.

Most researchers who have investigated changes in temperature and other properties of deep Atlantic water in the Arctic Basin correlate these changes with modification of atmospheric circulation (increase in its intensity and recurrence of cyclone penetration to high latitudes). An indirect confirmation of the increased inflow of Atlantic water to the Norwegian Sea and farther north is provided by the aforementioned decrease in surface water layer salinity in the eastern part of the sea as a result of adaptation of the field of masses to the system of currents. However, surface water freshening is accompanied by the increased stability of water masses, which leads to a decreased rate of deep-water cooling. As a result, the temperature of the water flowing to the Arctic Basin increases. This is confirmed by the results of isotopic analysis of the seabed (Duplessy, 1980; Flohn, 1980), showing that during glacial epochs, when vertical circulation was restricted by a

1900~ 1920 1940 1960 1980 2000


Figure 4.24. Long-term variability of Atlantic water temperature in the Arctic Basin in the twentieth century (Polyakov et al., 2005). Dashed line is extrapolation for missing data.

comparatively thin surface layer, the near-bottom water temperature in the Norwegian Sea was higher than during the current epoch.

To determine the influence of the distribution of water masses (dynamic heights) in the Arctic Basin on sea ice extent of the Asian-shelf Arctic Seas, Koltyshev and Timokhov (1997) used expansions of the fields of dynamic heights and sea ice extent by EOF and by singular values (SV). The cross-correlation analysis allowed them to determine that the sea ice extent of some marginal seas depends on the structure of dynamic height fields during the preceding significant time intervals (from several months to two years). An inverse impact of sea ice extent on water circulation in some seas was also revealed and attributed to dependence of surface water salinity on sea ice extent. Koltyshev and Timokhov (1997) speculate that sea ice and ocean changes are organized as a self-oscillating system with a period of approximately 6 years.

Was this article helpful?

0 0

Post a comment