New evidence of warming propagating toward the Arctic Ocean



[1] The dramatic changes in the Arctic climate observed in recent years have generated an urgent need to investigate the Arctic Ocean's heat budget. Mooring data and synoptic observations in the Fram Strait (FS) region have shown that increasing amounts of heat have been transported into the Arctic Ocean (AO) in recent times. Here we present results from observations conducted in the West Spitsbergen Current (WSC) in summers 2000–2005. The study was motivated by the strong warm anomalies seen in the Atlantic Water (AW) layer over a large area of the WSC, and changes in the WSC structure. We conclude that the warm signal was only approaching the FS, and we expect that high heat transport through the strait will continue and be even higher than during the last 6 years, mostly due to increasing activity and temperature of the western branch of WSC.

1. Introduction

[2] Warm, saline AW is the main source of heat for the AO. AW is transported into the Norwegian Sea by two branches of the Norwegian Atlantic Current (NwAC) [Poulain et al., 1996]. The eastern branch flows through the Faroe-Shetland Channel and continues along the Norwegian shelf edge as the Norwegian Atlantic Slope Current (NwASC). After passing Norway, the NwASC divides: one part flows east into the Barents Sea [Schauer et al., 2002], while the other part continues northward in the direction of the FS as the WSC [Aagaard et al., 1985]. The western branch of the NwAC is formed by the jet stream at the Polar Front [Orvik et al., 2001]; further to the north it continues over the Mohns and Knipovich Ridges, forming the western branch of the WSC.

[3] Two main streams of northward flow are recognizable in Figure 1: the core of the WSC continuing along the Barents Sea and Spitsbergen shelf break, and the colder, less saline stream of AW flowing over the submarine ridges system. Controlled by the bottom topography, both branches converge west of Spitsbergen [Piechura and Walczowski, 1995; Walczowski et al., 2005], only to divide again to produce the multi-path structure of the FS flows.

Figure 1.

2000–2005 – mean July temperatures and baroclinic currents at 100 m. The station grid is marked. The northern region is situated above line ‘A’.

[4] In recent years, the advection of warm anomalies along the path of the AW from the North Atlantic into the AO has been observed [Polyakov et al., 2005]. The core of the NwASC has warmed up by over 1°C during the last 10 years [Orvik and Skagseth, 2005]. An increase in the heat transport into the AO in 1997–2000 measured by an array of moorings in the FS was reported by Schauer et al., 2004. Modeling [Karcher et al., 2003; Zhang, 2005] has also shown that the rise of heat transported by AW may be the main reason for the AO warming.

[5] In this study we will concentrate on the variations in the heat content of the AW layer and the spatial structure of the WSC. Data collected during 6 summer cruises allow us to investigate the summer-to-summer variability of AW properties and pathways. An important and more general question is how can upstream measurements of the WSC help to predict changes in processes like heat transport through the FS. We argue that the low speed of signal propagation and the conservative nature of the tracers permit anticipating probable trends in heat transport into the Arctic Ocean.

2. Data and Methods

[6] The Institute of Oceanology of the Polish Academy of Sciences (IO PAS) has been studying the Nordic Seas since 1987, and the present grid of stations (Figure 1) has been covered every summer (June–July) since 2000 during cruises of r/v Oceania. A Seabird 9/11 CTD system was used for measurements. Vertical profiles of the temperature (T), salinity (S), and specific volume anomaly were averaged every 5 m. To obtain the horizontal distributions of these properties, the results were interpolated into rectangular grids using the kriging procedure. The grids were smoothed with a linear convolution low-pass filter. The heat content (with reference to −0.1°C) in the AW (S > 34.92 PSU and T > 0°C) layer was calculated. The mean temperature and salinity of the AW, AW layer thickness, heat content, and geostrophic currents referenced to 1000 dbar were calculated for each summer for the common area of about 297000 km2 (over 1100 CTD casts). The anomaly from the 6-year mean was also computed for each variable. Monthly temperature fields from Polar Science Center Hydrographic Climatology (PHC 3.0) [Steele et al., 2001] (Available at were compared to our observations.

3. Warm Anomalies in the West Spitsbergen Current

[7] Our data show that the mean temperature of the Nordic Seas in June–July 2000–2005 was much higher in comparison with climatology (Figure 2), with differences as high as 1.5°C. The summers 2000–2005 were warmer than average, but considerable year-to-year differences were observed as well. Time series of measurements along the 76°30′N parallel reveal a significant increase in the temperature and salinity of AW. T and S of AW, measured in summer at 200 m and averaged between longitude 009° and 012°E, have increased over 10 years by about 1°C and 0.06 PSU respectively, and have reached two minima (1997 and 2003) and two maxima (2001 and 2005) (Figure 3). The post-1997 high temperature and salinity correlate with the increase in heat transport through the FS reported by Schauer et al. [2004] for the 1997–2000. Since 2003, temperature and salinity increases also correlate with the higher heat flux through the FS recorded in 2003–2005 (A. Beszczynska-Möller, personal communication, 2005).

Figure 2.

Differences between the mean temperature in summers from 2000 to 2005 and July PHC 3.0 climatology data at 100 m.

Figure 3.

Temperature (thick line) and salinity at 200 m depth; section 76°30′N, averaged between longitude 009°–012°E.

[8] Horizontal distributions of T, S, heat content, and their anomalies have also revealed temporal variability in AW properties. The coldest summer was 2003 and the warmest was 2005. Moreover, the heat content anomalies in the AW layer (Figure 4) show that the heat transport was pulsating in nature. Positive heat anomalies usually had the structure of an anticyclonic eddy, whereas negative ones displayed a cyclonic circulation pattern. In summer 2005 the AW occupying the entire investigated region was unusually warm and saline. Two anticyclonic anomalies over the submarine ridges were especially intensive. The heat content in the AW layer has been increasing since 2001 (Figure 5), and since 2003 this rise has been very rapid. At the same time the volume of AW has decreased slightly, which means that the increasing heat content during the last 2 years was due to the higher temperature of AW.

Figure 4.

AW layer heat content anomalies and baroclinic current anomalies at 100 m. Anomalies calculated as differences from 2000–2005 July mean.

Figure 5.

Heat content (thick line) and volume of the AW layer for the entire study area.

[9] Geostrophic currents were also more intensive in 2005. Unsmoothed fields (Figure 6) reveal a large eddy at the latitude of 73°20′N and an eddy or frontal meander at 76°10′N in the western branch of the WSC. The southern eddy was 150 km in diameter, the AW reached a depth of 900 m, and the density structure was disturbed even below a depth of 2000 m. These large heat anomalies were carried by these structures. Intermittent acceleration of the along-slope branch was also observed. Both branches converged at 77°30–78°N, then diverged again downstream, forming a multi-path flow structure in the FS. Part of the AW flowing along the shelf break as the so-called Svalbard Branch entered the AO along the slope.

Figure 6.

Temperature and baroclinic currents at 100 m. Non-filtered grid. Summer 2005.

4. Concluding Remarks

[10] In 2005 a substantial rise in the salinity and temperature of the AW was observed (Figures 3 and 7) . Since salinity is a more conservative property than temperature, its increase suggests the advective character of the observed anomalies, independent of the seasonal variations. Furthermore, since the speed of signal propagation is low, integrated values like heat content from synoptic observations conducted over such a large area can be used to predict tendencies in the heat flux entering the Arctic Ocean.

Figure 7.

Maximum AW salinity and baroclinic currents at 100 m in the Fram Strait region.

[11] Estimations of speed of signal propagation in the Nordic Seas are sparse; various sources report 2–3.2 cm/s [Furevik, 2000]. The mean velocity of 3.8 cm/s was derived from the time lag between the warm signal observed at the Svinoy Section (63°N) and in FS (78°55′N) [Polyakov et al., 2005]. The advection of the anomaly along the distance of 900 km from the southern side of investigated region to the FS would take 290 days at a mean speed of 3.8 cm/s and 520 days at 2 cm/s.

[12] In 1997–2000 the rising heat transport into the AO was caused partly by temperature increase and partly by stronger flow [Schauer et al., 2004]. The Nordic Seas in 2004 were considered very warm [Working Group on Oceanic Hydrography, 2005], and 2005 seems to have been even warmer. The increase in the heat content in the northern part of the study area (see Figure 1) was considerable at from 3.46·1020 J to 3.85·1020 J. It was caused mostly by the increase in the mean AW temperature from 3.08°C to 3.40°C, whereas the AW volume in this area increased only slightly, from 28060 to 28426 km3.

[13] Thus, we can assume that the warming of AW flowing into the AO will continue, probably with even higher intensity in late 2005 and early 2006. Assuming the 3 cm/s of mean advection speed, observed in 2005 the northern anomaly that carried about 2.4·1020 J of heat may produce 47 TW of heat flux during two months. The complex flow structure in the Fram Strait, mostly the recirculation of AW, may cause a large part of this heat to be recirculated westward into the East Greenland Current and then southward. On the other hand, it was observed that recirculation was intensive in cold years like 2003, and inflow into the AO was intensive during warm summers (Figure 7).

[14] Orvik and Skagseth [2003] have shown that the interannual variability of the NwASC is linked with the wind stress curl in the North Atlantic 15 months earlier. Therefore, the NwASC should have decreased in 2005 because of the decrease in the 2004 wind stress curl at 55°N [Drange et al., 2005], but the 2004 and 2005 heat content anomaly (Figure 4) show that the NwASC may not be the major source of the last AW warming and that the western branch of the NwAC and the WSC could also be playing a very significant part.

[15] Our results show that heat transport has a pulsating nature and that the spatial structure of currents varies. There are years when the eastern branch is dominant (2002 in the entire study area, 2004 in the northern part), but the current picture in 2005 was really quite different with substantial transport by the western branch coupled with intensive transport along the shelf break. Moreover, the circulation pattern in the FS differed from the usual one. In warm years, like 2004, AW inflow to the AO was concentrated along the shelf break in the Svalbard Branch (Figure 7). In 2005, however, the high temperature and salinity signal was shifted slightly to the west and was seen in a broad area. There was a pronounced Svalbard Branch inflow, as well as a northward flow along the western slope of the Yermak Plateau.

[16] When the anomalies observed in summer 2005 at 76°–77°N and 73°–74°N pass through the FS, they will carry large amounts of heat into the AO. In the future, the Svalbard Branch inflow should become apparent at existing and planned current meter arrays at the slope of the eastern Eurasian Basin. The pathway of the Yermak Plateau branch is not so clear; it will probably continue along the edge of the plateau, turn east, and join the Svalbard Branch.

[17] In this paper we have concentrated on the temperature and heat content of the AW layer, but a significant rise in AW salinity has been observed as well (Figures 3 and 7). At a time when the freshwater outflow from the AO into the Nordic Seas intensifies, salt input into this dominant source area for the North Atlantic Thermohaline Circulation may be very important for the maintenance of Thermohaline Circulation.


[18] We are grateful to the entire r/v Oceania crew who participated in the work at sea. Special acknowledgement goes to Agnieszka Beszczynska-Möller for information and discussion.

[19] This research has been supported by grant from the European Union Fifth Framework Program project ASOF-N (Arctic-Subarctic Ocean Flux Array for European Climate: North), contract number EVK2-CT-200200139.