Abstract

A potential for climate predictability is rooted in anomalous ocean heat transport and its consequent influence on the atmosphere above. Here the propagation, drivers, and atmospheric impact of heat anomalies within the northernmost limb of the Atlantic meridional overturning circulation are assessed using a multicentury climate model simulation. Consistent with observation-based inferences, simulated heat anomalies propagate from the eastern subpolar North Atlantic into and through the Nordic seas. The dominant time scale of associated climate variability in the northern seas is 14 years, including that of observed sea surface temperature and modeled ocean heat content, air–sea heat flux, and surface air temperature. A heat budget analysis reveals that simulated ocean heat content anomalies are driven by poleward ocean heat transport, primarily related to variable volume transport. The ocean’s influence on the atmosphere, and hence regional climate, is manifested in the model by anomalous ocean heat convergence driving subsequent changes in surface heat fluxes and surface air temperature. The documented northward propagation of thermohaline anomalies in the northern seas and their consequent imprint on the regional atmosphere—including the existence of a common decadal time scale of variability—detail a key aspect of eventual climate predictability.

1. Introduction

The Atlantic region exhibits distinct interannual to multidecadal variability (Deser and Blackmon 1993; Kushnir 1994; Frankcombe et al. 2010; Williams et al. 2014), reflected in upper-ocean thermohaline anomalies that propagate persistently through the North Atlantic Ocean and Nordic seas toward the Arctic (Sutton and Allen 1997; Polyakov et al. 2005; Holliday et al. 2008). Decadal variability in ocean temperature plays an important role in the marine climate system (e.g., Drinkwater et al. 2014), influencing marine life from primary production to cod stocks (Helland-Hansen and Nansen 1909; Hátún et al. 2009). Ocean heat anomalies also play an important role in Arctic sea ice variability (Francis and Hunter 2007; Årthun et al. 2012; Onarheim et al. 2014; Carmack et al. 2015), which in turn could influence weather conditions and climate (e.g., Screen et al. 2014; Vihma 2014).

Anomalous ocean heat can extend its influence beyond the marine climate by being imprinted on the atmosphere (Rhines et al. 2008; Farneti and Vallis 2011; Gulev et al. 2013; Schlichtholz 2013), acting to increase the persistence of atmospheric circulation anomalies and, hence, provide predictability of atmospheric variability and continental climate (e.g., Sutton and Hodson 2005). This, however, requires that oceanic variability is communicated to the atmosphere through surface heat fluxes. Understanding the mechanisms and time scales involved in the propagation of ocean heat anomalies and how they interact with the atmosphere is thus a prerequisite for skillful climate predictions for the North Atlantic/Arctic sector (Latif and Keenlyside 2011; Meehl et al. 2014).

The flow of warm, saline Atlantic waters toward higher latitudes takes place with the North Atlantic Current and its poleward extension, the Norwegian Atlantic Current (NwAC; Fig. 1a). In the Nordic seas, the NwAC consists of two branches; a western branch enters the Nordic seas over the Faroe–Iceland Ridge and is topographically guided northward along the front between the Arctic and Atlantic waters, while a warmer and more saline eastern branch inflows through the Faroe–Shetland Channel and continues north as a near-barotropic shelf-edge current (Orvik et al. 2001). Upon reaching the western boundary of the Barents Sea the eastern branch of the NwAC bifurcates flowing eastward into the Barents Sea, while the northward flow converges with the western NwAC and continues toward the Fram Strait as the West Spitsbergen Current. Part of the West Spitsbergen Current continues north into the Arctic Ocean (Rudels et al. 1999), but the majority of the current recirculates westward in the Fram Strait (Bourke et al. 1988) and joins the southward flowing deeper branch of the East Greenland Current en route to the Denmark Strait, thus forming a cyclonic loop within the Nordic seas. While traversing the periphery of the Nordic seas and the Arctic Ocean, the warm and saline Atlantic water is gradually transformed into a colder and fresher outflow as a result of oceanic heat loss and freshwater input (Mauritzen 1996; Rudels et al. 1999). Following Eldevik et al. (2014), the three regions connected by the NwAC (Fig. 1a)—the northern North Atlantic, the Nordic seas, and the Arctic Ocean—will hereafter be jointly referred to as the northern seas.

Fig. 1.

(a) Observed and (b) modeled winter (December–April) upper-ocean temperature (shading) and sea ice extent (white line; defined where the sea ice concentration is 15%) in the northern seas. Observations are from HadISST (Rayner et al. 2003). The arrows indicate the main features of the near-surface circulation (NwAC: Norwegian Atlantic Current; WSC: West Spitsbergen Current; EGC: East Greenland Current). Modeled temperature and velocities are averages within the surface mixed layer (note the different velocity scales). Isobaths are given every 1000 m (thin gray lines) and for 500-m depth (black line), which roughly marks the continental slopes. DS: Denmark Strait; RT: Rockall Trough; Sh: Shetland.

Fig. 1.

(a) Observed and (b) modeled winter (December–April) upper-ocean temperature (shading) and sea ice extent (white line; defined where the sea ice concentration is 15%) in the northern seas. Observations are from HadISST (Rayner et al. 2003). The arrows indicate the main features of the near-surface circulation (NwAC: Norwegian Atlantic Current; WSC: West Spitsbergen Current; EGC: East Greenland Current). Modeled temperature and velocities are averages within the surface mixed layer (note the different velocity scales). Isobaths are given every 1000 m (thin gray lines) and for 500-m depth (black line), which roughly marks the continental slopes. DS: Denmark Strait; RT: Rockall Trough; Sh: Shetland.

Temperature anomalies have been observed to propagate northward from the eastern subpolar North Atlantic along the path of the North Atlantic Current and NwAC (Furevik 2000; Holliday et al. 2008; Chepurin and Carton 2012; Yashayaev and Seidov 2015). In the northern seas, anomalies travel from the Greenland–Scotland Ridge to the west coast of Svalbard in approximately 1–3 years (Dickson et al. 1988; Eldevik et al. 2009). This corresponds to a propagation speed of 2–5 cm s−1, which is an order of magnitude less than the typical current speed of the NwAC (Orvik et al. 2001). Both anomalous air–sea heat fluxes due to changes in the large-scale atmospheric circulation (Furevik and Nilsen 2005) and the changing composition and strength of ocean currents have been suggested to generate temperature anomalies in the northern seas (e.g., Furevik 2001; Carton et al. 2011; Mork et al. 2014). Specifically, Mork et al. (2014) found that heat fluxes explain about half of the observed interannual heat content variability, but that the fraction varies considerably in time. Carton et al. (2011), on the other hand, found that surface heat flux variations in some cases act to reinforce anomalies, but that the contribution was too small to explain the concomitant changes in ocean heat storage. It is in most cases, however, not possible to construct a closed observation-based heat budget because of sparse data coverage, especially in terms of ocean current measurements. The relative importance of ocean and atmosphere in modifying ocean heat anomalies can therefore not be fully distinguished from observations. Heat flux reanalysis products also partly disagree, making it sometimes problematic to compare with changes in observed hydrography (Carton et al. 2011).

The purpose of this paper is twofold: to disentangle the contributions from ocean circulation and air–sea exchange in the propagation of ocean heat anomalies from the North Atlantic toward the Arctic, and to assess the potentially predictable relation between anomalous ocean heat and climate in the northern seas region. To this end, a 500-yr control simulation with the Bergen Climate Model is used (Otterå et al. 2009). The model analysis is aided by historical sea surface temperatures (HadISST; Rayner et al. 2003).

The model and the observations are introduced in section 2. In section 3 we evaluate the model performance for the northern seas. The propagation and drivers of anomalies are then analyzed in section 4 and section 5, while the link to upstream variability in the subpolar North Atlantic is discussed in section 6. The atmospheric imprint of ocean heat anomalies and the identified characteristic time scale of oceanic and atmospheric variability are discussed in section 7. Finally, the main conclusions and implications are presented in section 8.

2. Data and methods

a. Bergen Climate Model

This study uses a 500-yr preindustrial control simulation from the Bergen Climate Model (BCM), a fully coupled atmosphere–ocean–ice general circulation model. A general description of the model is given by Furevik et al. (2003), while the model run used is described in Otterå et al. (2009). Only a short summary is given here.

The ocean component of BCM is a modified version of the Miami Isopycnic Coordinate Ocean Model (MICOM; Bleck et al. 1992). The version used for this model run uses potential density with reference pressure at 2000 dbar as vertical coordinates ( coordinates). The ocean model consists of 34 isopycnic layers, ranging from kg m−3 to kg m−3, below a nonisopycnic mixed layer. The mixed layer depth is calculated from the turbulent kinetic energy balance of a Kraus–Turner-type one-dimensional mixed layer model (Gaspar 1988), with modifications detailed in Medhaug et al. (2012). The horizontal grid resolution is 2.4° longitude × 0.8° latitude at the equator, becoming more isotropic with increasing latitude. In the northern seas the horizontal resolution ranges from 70 to 100 km. MICOM is coupled to a multicategory dynamic–thermodynamic sea ice model, Global Experimental Leads and Ice for Atmosphere and Ocean (GELATO; Salas-Mélia 2002). The atmospheric component is ARPEGE-CLIMAT3 (Déqué et al. 1994), a low-top spectral model with a horizontal resolution of ~2.8° and 31 vertical levels. Fluxes of mass, energy, and/or momentum are calculated in ARPEGE and communicated to the ocean via the OASIS (Terray and Thual 1995) coupler. External forcing from, for example, solar insolation and greenhouse gases, is set to constant preindustrial values. No flux corrections are applied, allowing the model to freely develop its own climatology. The initial conditions for the preindustrial control simulation are obtained from the end of a 500-yr spinup integration [detailed in Otterå et al. (2009)], after which the model is run for 600 years. Here we use the last 500 years of the simulation.

b. Methods

Ocean heat anomalies are assessed as follows. The heat content is calculated as

 
formula

where T is temperature, is density of seawater, and is the specific heat capacity of seawater. The heat content is calculated between the free surface η and the full depth of the ocean, D, using a reference temperature C, although we note that heat content anomalies (relative to the local mean) presented herein are practically insensitive to the specific reference temperature (not shown). We also note that an ocean heat budget calculated from the heat convergence of a closed mass budget (section 5) is independent of . Monthly anomalies at each grid point are then obtained by subtracting the respective mean seasonal cycle. Unless stated otherwise time series are then low-pass-filtered with a third-order Butterworth filter with a cutoff period of 3 years to emphasize interannual to decadal variability.

Statistical significance is assessed using a two-tailed Student’s t test, adjusted for serial autocorrelation (Chelton 1983). All correlations given in the text are significant at the 95% confidence level.

c. Complex principal component analysis

Propagating phenomena can be identified and objectively analyzed from a complex principal component (CPC) analysis that detects traveling waves in the input time series. A full description of the procedure is given by Horel (1984), and only a short summary is presented here. First, a complex dataset is formed from the original data (by rotating its Fourier components π/2). Complex eigenvectors are then computed from the cross-covariance matrix derived from the complex dataset. From the covariance matrix, the CPCs [] and the complex empirical orthogonal functions [CEOFs; ] are calculated. The (complex) dataset can thus be represented as a sum of the contribution from the N principal components:

 
formula

where x denotes spatial position and t is time. The asterisk denotes complex conjugation. The elements of the CPC time series can furthermore be written in the form of an amplitude and a phase ; . The importance of each CEOF (mode) is defined as the proportion of variance explained by each principal component.

d. Observed SST

To corroborate the model analysis we use sea surface temperature (SST) data from the Hadley Centre (HadISST; Rayner et al. 2003) covering the period between 1870 and 2013. These data have a spatial resolution of 1° longitude by 1° latitude and monthly temporal resolution. We will only consider winter (December–April) SST as it represents the upper-ocean heat content as a result of a deep winter mixed layer (Nilsen and Falck 2006). The observation-based analysis is furthermore restricted to the southern Norwegian Sea to avoid the potential influence of sea ice. A good agreement in terms of interannual to decadal variability has been found between HadISST and data from standard hydrographic sections in the northern seas (Hughes et al. 2009).

3. Model performance in the northern seas

To assess the propagation of ocean heat anomalies in the northern seas it is essential that the model of choice is able to adequately represent the northward flow of Atlantic water and the gradual transformation into dense overflow water as it circulates the periphery of the Nordic seas. Previous applications of the Bergen Climate Model in the northern seas (e.g., Otterå et al. 2010; Langehaug et al. 2012a,b; Medhaug et al. 2012; Lohmann et al. 2014) have found that the model realistically simulates the structure and the mean poleward heat transport of the NwAC, as well as the dense overflow and fresh surface waters in the Denmark Strait. Notably, the thermohaline contrast between these three water masses occupying the Greenland–Scotland ridge is consistent with observations, although the model hydrography is skewed toward warmer and more saline properties. For the Atlantic inflow specifically, the model is ~2°C warmer and ~0.5 psu saltier than observations (cf. Fig. 7 in Langehaug et al. 2012a). The associated modeled volume transport into the Nordic seas is 7.4 Sv for the NwAC, 2.1 Sv in the East Greenland Current, and 5.7 Sv of overflow water, which is in good agreement with observational estimates of 8.5, 0.4–2.1, and 6.4 Sv, respectively [see Langehaug et al. (2012a) and references therein; 1 Sverdrup (Sv) ≡ 106 m3 s−1).]. The model also captures the bifurcation at the western boundary of the Barents Sea (Fig. 1b), with a heat transport into the Barents Sea (61 TW; 1 TW = 1012 J s−1; Medhaug et al. 2012), which is close to that observed (Årthun et al. 2012). The realistic model transports and properties of inflowing and outflowing waters point to an accurate modification of water masses within the Nordic seas. This is corroborated by Langehaug et al. (2012b), who found a realistic structure of surface-forced water mass transformation (diagnosed from surface buoyancy fluxes) along the path of the North Atlantic Current in the model.

The observed and simulated winter sea ice extent is shown in Fig. 1. The model ice cover is in good agreement with observations in the Norwegian Sea and western Barents Sea, while it is generally larger than the observed in the Greenland Sea. However, as noted by Smedsrud et al. (2013), a more extensive ice cover in the model is reasonable as the simulation uses constant preindustrial external forcing and therefore does not include recent sea ice decline in the Arctic. The simulated minimum and maximum winter sea ice edge are furthermore in agreement with the observed minimum and maximum (Smedsrud et al. 2013); that is, the model variability is within the observed range.

Admittedly, the Bergen Climate Model and other global climate models are unable to resolve the smaller-scale features of ocean circulation (e.g., mesoscale ocean eddies and narrow boundary currents). However, multidecadal variability of the coupled atmosphere–ocean system can only be studied using relatively coarse climate models, as multicentury simulations with eddy-resolving grid resolution are generally not available. Fully coupled climate models are nevertheless valuable and necessary tools for assessing climate variability on decadal time scales and beyond.

4. Propagation of simulated ocean heat anomalies

To assess the properties and modification of modeled ocean heat anomalies in the northern seas we first need to determine the path of propagation. Based on the mean circulation (illustrated by the barotropic stream lines in Fig. 2) and extent of the simulated NwAC (Fig. 1b), 11 stations (St1–St11) have been defined that capture the mean propagation. This includes one station in the North Atlantic, corresponding to the Rockall Trough, one at the Greenland–Scotland ridge, and nine stations downstream within the Nordic seas. The grid points included in each station are shown as circles in Fig. 2. The lagged correlation between adjacent stations is generally high (>0.7; lags vary, but are typically around 6 months) for both heat (Fig. 2) and salt content (not shown), except for lower values between St2 and St3, and St9–St11. The former could be a result of both variable communication between the North Atlantic and Nordic seas (e.g., Hátún et al. 2005) and hydrographic variability internal to the Nordic seas (e.g., Mork et al. 2014), while the less coherent signal between St9 and St11 could be a result of the branch of Atlantic water entering the Nordic seas west of Iceland (Fig. 1b; Langehaug et al. 2012a) or branching of the southward flow in the south Greenland Sea (Mauritzen 1996).

Fig. 2.

Barotropic streamlines showing the mean cyclonic circulation within the northern seas [thick black lines; plotted for −1.5, −2, and −3 Sv; 1 Sverdrup (Sv) ≡ 106 m3 s−1]. Isobaths are given for 500-, 2000-, and 3000-m depth (gray lines). The colored circles show the grid points used to define the path of ocean heat anomalies, where the associated color indicates the maximum lagged correlation between adjacent stations; i.e., the correlation shown for St2 corresponds to r(St1,St2). Station numbers are assigned and used in the text.

Fig. 2.

Barotropic streamlines showing the mean cyclonic circulation within the northern seas [thick black lines; plotted for −1.5, −2, and −3 Sv; 1 Sverdrup (Sv) ≡ 106 m3 s−1]. Isobaths are given for 500-, 2000-, and 3000-m depth (gray lines). The colored circles show the grid points used to define the path of ocean heat anomalies, where the associated color indicates the maximum lagged correlation between adjacent stations; i.e., the correlation shown for St2 corresponds to r(St1,St2). Station numbers are assigned and used in the text.

The time evolution of ocean heat anomalies along the defined path is shown in Fig. 3a. Propagating warm and cold anomalies are evident throughout the time period. Heat anomalies with a depth-averaged standard deviation of 1.4 × 1019 J (corresponding to 0.5°C) are associated with salinity anomalies of 0.05, warm conditions being accompanied by higher salinities (Fig. 3a) (i.e., anomalies being largely density compensated). Elevated variability is found in the southern end of the Norwegian Sea (St2–St3) and in the area between Norway and Svalbard (St6–St7). This spatial pattern is consistent with that inferred from observed Norwegian Sea heat content variability (Skagseth and Mork 2012). The observed anomalies discussed by Furevik (2001) also showed the largest amplitude in the Sørkapp section at 76.5°N (approximate position of St7). The along-path evolution (relative to the local mean) is determined by the concomitant anomalous forcing (ocean and atmosphere) within the northern seas. A northward strengthening of an anomaly can for instance be explained by anomalously low surface heat loss or by an increased advection speed (Furevik 2001).

Fig. 3.

(a) Temporal development of low-pass-filtered heat content (shading; units in 1019 J) and depth-averaged salinity anomalies (contours; plotted for 0 and ±0.05, dashed lines corresponding to a negative salinity anomaly) at defined stations (averages over grid points shown in Fig. 2). The thick dashed line corresponds to a propagation speed of 2 cm s−1. (b) The dominant mode of propagation derived from complex principal component analysis (units in 0.5 × standard deviation), explaining 50% of the total variance. (c),(d) Phase angle as a function of station number and time, respectively. Increasing phase angle with increasing distance (station) corresponds to a cyclonic propagation. The dashed line in (c) corresponds to a propagation speed of 2 cm s−1, whereas in (d) it corresponds to constant phase propagation.

Fig. 3.

(a) Temporal development of low-pass-filtered heat content (shading; units in 1019 J) and depth-averaged salinity anomalies (contours; plotted for 0 and ±0.05, dashed lines corresponding to a negative salinity anomaly) at defined stations (averages over grid points shown in Fig. 2). The thick dashed line corresponds to a propagation speed of 2 cm s−1. (b) The dominant mode of propagation derived from complex principal component analysis (units in 0.5 × standard deviation), explaining 50% of the total variance. (c),(d) Phase angle as a function of station number and time, respectively. Increasing phase angle with increasing distance (station) corresponds to a cyclonic propagation. The dashed line in (c) corresponds to a propagation speed of 2 cm s−1, whereas in (d) it corresponds to constant phase propagation.

In the North Atlantic and southern Norwegian Sea (St1–St5) the magnitude of ocean heat content variability is largest between 300- and 500-m depth (Fig. 4). The stations farther downstream (St6–St9) show maximum variability deeper in the water column (although less pronounced at St8–St9), reflecting the along path modification and deepening of the Atlantic water. The marked change between St7 and St8 relates to the boundary between the ice-free Norwegian Sea and the seasonally ice covered Fram Strait and Greenland Sea (Fig. 1b), that is, the northward extent of the Atlantic domain. The associated temperature variability is approximately 0.6°–1.0°C in the Norwegian Sea, while it is smaller (<0.4°C) in the Greenland Sea and in the Rockall Trough. The surface intensified variability at St6 and St7 is most likely related to large surface heat loss (both modeled and observation-based; Langehaug et al. 2012b), whereas the large surface variations at St11 is caused by the Atlantic inflow west of Iceland. The depth of maximum heat content variability at St1–St6 corresponds the average depth of the winter mixed layer (Fig. 4). Noting that the depth of the winter mixed layer in the Norwegian Sea reflects the base of the Atlantic layer (Nilsen and Falck 2006), this suggests that heat content changes are largely driven by changes in the layer thickness and, hence, volume of Atlantic water (Sandø et al. 2012). This will be elucidated further in the next section.

Fig. 4.

Heat content (shading) and temperature (contours) variability (expressed as one standard deviation) at defined stations as a function of depth. The black line displays the average depth of the winter (December–April) mixed layer.

Fig. 4.

Heat content (shading) and temperature (contours) variability (expressed as one standard deviation) at defined stations as a function of depth. The black line displays the average depth of the winter (December–April) mixed layer.

Propagating phenomena can be objectively assessed from a complex principal component analysis, which detects traveling waves in the input time series and orders the dataset into modes of phase propagation in space and time according to the variance explained (section 2). The leading mode of phase propagation of along-path heat anomalies (Fig. 3b) explains 50% of the total variance in the full dataset (Fig. 3a) and is well separated from the second mode (accounting for 21% of the variance). The phase angle of the leading mode increases with increasing station number (Fig. 3c), which implies that the simulated heat anomalies predominantly travel along the rim of the basin in the direction of the mean current, consistent with observation-based inferences (Holliday et al. 2008; Eldevik et al. 2009). The 500-yr time period consists of 31 complete cycles with a phase propagation that is rather constant in time (Fig. 3d). This yields a period of 16 years. The circulation from St1 to St11 constitutes just over half a cycle, which, with a travel distance of about 4800 km, implies that the speed of modeled anomalies is on average 2 cm s−1, an estimate that is in reasonable agreement with observations (Furevik 2000; Polyakov et al. 2005; Holliday et al. 2008; Chepurin and Carton 2012).

The representativeness of the time scale associated with propagating anomalies obtained from the complex principal component analysis compared with the full variance can be evaluated by a frequency analysis of the anomalous heat content at individual stations (note that no filter is applied in the frequency analysis). Heat anomalies in the northern North Atlantic (St1) and the Norwegian Sea (St3) both have a significant (95% confidence level) characteristic time scale of 14 years (Fig. 5a). The 14-yr time scale is also clearly identifiable for salinity (Fig. 5b). All Atlantic-dominated stations (St1–St7; cf. Fig. 4) as well as downstream St8 and St9 show significant power on this time scale. As mentioned above, the weaker signature of a propagating signal at St10 and St11 is most likely a result of thermohaline anomalies exiting the Nordic seas both west (St11) and east of Iceland (Mauritzen 1996; Eldevik et al. 2009). Significant interdecadal variability is also found in the observation-based HadISST winter sea surface temperatures between 1870 and 2013 for the same region (60.5°–71.5°N, 0.5°–17.5°E; Fig. 5c), increasing the confidence in the model’s ability to simulate climate variability in the northern seas. The modeled ocean heat content also displays significant multidecadal variations (~40–50 yr). Variability on this time scale, however, will not be addressed here and the reader is referred to, for example, Frankcombe et al. (2010) and references therein for a discussion on mechanisms for multidecadal variability in the North Atlantic.

Fig. 5.

Power spectrum (thick lines) for (a) heat content and (b) salinity anomalies at St1 and St3 based on unfiltered data together with the theoretical red noise spectrum (thin solid lines) computed by fitting a first-order autoregressive process with a 95% confidence interval (thin dashed lines) around the red noise. (c) Power spectrum for observed (HadISST) winter (December–April) sea surface temperatures in the Norwegian Sea (60.5°–71.5°N, 0.5°–17.5°E) between 1870 and 2013. Winter SST is considered because it represents the upper-ocean heat content as a result of a deep winter mixed layer.

Fig. 5.

Power spectrum (thick lines) for (a) heat content and (b) salinity anomalies at St1 and St3 based on unfiltered data together with the theoretical red noise spectrum (thin solid lines) computed by fitting a first-order autoregressive process with a 95% confidence interval (thin dashed lines) around the red noise. (c) Power spectrum for observed (HadISST) winter (December–April) sea surface temperatures in the Norwegian Sea (60.5°–71.5°N, 0.5°–17.5°E) between 1870 and 2013. Winter SST is considered because it represents the upper-ocean heat content as a result of a deep winter mixed layer.

5. Heat budget for the Norwegian Atlantic Current

To assess the relative roles of ocean advection and air–sea fluxes in driving ocean heat anomalies, and how anomalous ocean heat might imprint on the atmosphere, the depth-integrated heat budget for the Norwegian Sea (Fig. 6) is now assessed in particular. The chosen area corresponds to the Atlantic-dominated eastern Nordic seas (Fig. 1) where the heat content variability is highest (Fig. 6) and interaction with the atmosphere is strongest (Langehaug et al. 2012b). The heat budget area is also similar to that used in the observation-based heat content analysis by Carton et al. (2011) and Mork et al. (2014). Although heat content variability in the whole water column is considered, changes in Norwegian Sea heat content predominantly reflect variability within the well-mixed Atlantic layer (; Nilsen and Falck 2006; Chepurin and Carton 2012), which is in contact with the atmosphere.

Fig. 6.

(a) Heat budget for the Norwegian Sea (low pass filtered), where is the heat content tendency, is the heat transport convergence, is the net surface heat flux (positive out of the ocean), and is a residual term. The domain (upper map) is bounded in the south by the Iceland–Scotland Ridge and in the west and north by the extent of the Atlantic domain, visible by the gradient in heat content variability (expressed as one standard deviation; units are 108 J m−2). The lower map displays the grid points used in the analysis. Colored circles and black dots present the ocean grid, while the red open squares are the atmospheric grid points used to analyze surface air temperature. (b) Heat transport anomalies (HT; low pass filtered) through the vertical sections of the control volume. The line colors correspond to the section color given in the lower map and the names refer to geographic location or major currents (: Atlantic water; : North Sea; : Greenland Sea; : Barents Sea; : West Spitsbergen Current).

Fig. 6.

(a) Heat budget for the Norwegian Sea (low pass filtered), where is the heat content tendency, is the heat transport convergence, is the net surface heat flux (positive out of the ocean), and is a residual term. The domain (upper map) is bounded in the south by the Iceland–Scotland Ridge and in the west and north by the extent of the Atlantic domain, visible by the gradient in heat content variability (expressed as one standard deviation; units are 108 J m−2). The lower map displays the grid points used in the analysis. Colored circles and black dots present the ocean grid, while the red open squares are the atmospheric grid points used to analyze surface air temperature. (b) Heat transport anomalies (HT; low pass filtered) through the vertical sections of the control volume. The line colors correspond to the section color given in the lower map and the names refer to geographic location or major currents (: Atlantic water; : North Sea; : Greenland Sea; : Barents Sea; : West Spitsbergen Current).

Changes in heat content within a control volume occur as a result of the imbalance between the surface heat flux (area A) and advective and diffusive heat transports through the vertical boundaries (area S):

 
formula

where is the heat content tendency, υ is the cross-sectional velocity into the control volume used to calculate the heat transport convergence (), q is the net ocean–atmosphere heat flux (the sum of both turbulent and radiative components) that integrated over an area A yields , and is a residual term representing the ocean heat transport into the domain resulting from parameterized diapycnal mixing and lateral turbulent mixing (Otterå et al. 2009). Note that positive surface heat fluxes indicate ocean heat loss. In Fig. 6a the time-varying heat budget components are plotted as anomalies (for presentation purposes only the first 100 years are plotted, while all calculations are based on the full 500-yr time series), calculated as previously described in relation to Eq. (1). The heat content rate of change is strongly associated with ocean heat transport convergence, both in terms of variability () and amplitude. The larger contribution from oceanic variability in driving heat content change is generally the case within the northern seas (Fig. 7) and especially along the path of the NwAC, whereas air–sea fluxes are relatively more important in parts of the Greenland Sea and in the northwestern Barents Sea.

Fig. 7.

Relative magnitude of oceanic () and atmospheric () heat budget terms, calculated as std()/std() and presented on a log scale. Values >1 indicate that the oceanic contribution is the largest.

Fig. 7.

Relative magnitude of oceanic () and atmospheric () heat budget terms, calculated as std()/std() and presented on a log scale. Values >1 indicate that the oceanic contribution is the largest.

The advective heat budget for the Norwegian Sea is in turn dominated by anomalous heat transport between Iceland and Scotland ()—that is, the Atlantic inflow (; Fig. 6b). The heat transport anomalies associated with the Atlantic inflow lead the northern () and eastern () outflow by about 2.5 years, consistent with the calculated speed of anomalies. Heat transport anomalies [] can occur as a result of changes in advection speed () or temperature (), or through eddy fluxes (). The respective contributions of these terms to the Iceland–Scotland heat transport are shown in Fig. 8. The variability of is dominated by , that is, velocity and thus volume transport fluctuations rather than temperature fluctuations drive heat transport anomalies into the area. The dominant role of ocean advection on heat storage rates on interannual to decadal time scales agrees with recent findings from the subpolar North Atlantic (Buckley et al. 2014; Williams et al. 2014).

Fig. 8.

Heat transport anomalies (low pass filtered) between Iceland and Scotland (Fig. 6; ) decomposed into a temperature component (), a velocity component (), and an eddy component ().

Fig. 8.

Heat transport anomalies (low pass filtered) between Iceland and Scotland (Fig. 6; ) decomposed into a temperature component (), a velocity component (), and an eddy component ().

Heat anomalies in the Norwegian Sea (Fig. 6), rooted in the ocean, are furthermore found to force changes in air–sea fluxes. The correlation between ocean heat transport convergence and surface heat loss is 0.66 (i.e., anomalously high ocean heat transport corresponds to enhanced heat loss to the atmosphere). Consistent with variable air–sea exchange associated with ocean heat anomalies, the surface heat flux and surface air temperature (SAT) within the Norwegian Sea (evaluated at grid points shown in Fig. 6) also have the same decadal-scale oscillation of 14 years (Fig. 9a). The atmospheric temperature anomalies covary with ocean heat content (), leading to a reduced thermal contrast between ocean and atmosphere and thus contributing to weaker air–sea fluxes and reduced damping of ocean heat anomalies []. Changes in surface air temperature lag variations in the Atlantic inflow () by 2–3 years (). The potential predictability of atmospheric variability from ocean heat anomalies is further discussed in section 7.

Fig. 9.

Power spectrum for (a) surface heat flux () and surface air temperature (SAT) in the Norwegian Sea (cf. Fig. 6), and (b) the heat transport between Iceland and Scotland () and the strength of the subpolar gyre (SPG) based on unfiltered data. The red noise spectrums and 95% confidence intervals are also plotted (see Fig. 5 for description). The SPG strength is calculated as the absolute value of the minimum barotropic streamfunction in the subpolar region (Langehaug et al. 2012a).

Fig. 9.

Power spectrum for (a) surface heat flux () and surface air temperature (SAT) in the Norwegian Sea (cf. Fig. 6), and (b) the heat transport between Iceland and Scotland () and the strength of the subpolar gyre (SPG) based on unfiltered data. The red noise spectrums and 95% confidence intervals are also plotted (see Fig. 5 for description). The SPG strength is calculated as the absolute value of the minimum barotropic streamfunction in the subpolar region (Langehaug et al. 2012a).

6. Source of northern sea heat anomalies

The temporal development of heat content anomalies in the northern seas (Fig. 3) implies a propagating signal. The poleward progression of thermohaline anomalies—heat and freshwater—is also a robust finding in observations and in other ocean and climate models (e.g., Dickson et al. 1988; Hansen and Bezdek 1996; Krahmann et al. 2001; Holliday et al. 2008; Chepurin and Carton 2012; Glessmer et al. 2014). There is nevertheless no complete mechanistic understanding of the driving mechanisms of these anomalies, including the roles of ocean dynamics and stochastic atmospheric forcing [see review by Liu (2012)]. In support of the latter, a number of studies have related low-frequency temperature variability in the North Atlantic to atmospheric variability associated with the winter North Atlantic Oscillation (NAO; Hurrell 1995). Visbeck et al. (1998) and Krahmann et al. (2001) demonstrated how the formation and propagation of temperature anomalies along the pathway of the North Atlantic Current can be obtained by temporal changes in NAO-like wind forcing. The propagation was found to be a result of both advection of existing temperature anomalies by the mean ocean currents and locally generated anomalies from spatial variations in the external forcing. Simple advection of coherent temperature anomalies through the North Atlantic is also not supported by recent drifter studies (e.g., Burkholder and Lozier 2014), showing no direct advective pathway of anomalous heat between the subtropical and subpolar region.

Variable (NAO like) atmospheric forcing can also induce upper-ocean temperature anomalies through modulation of the North Atlantic Ocean circulation and subpolar gyre (SPG) strength, driving changes in poleward heat transport (Czaja and Marshall 2001; Eden and Jung 2001; Lohmann et al. 2009). This has also been shown for the Bergen Climate Model (Langehaug et al. 2012a; Medhaug et al. 2012). It has previously been demonstrated both from observations (e.g., Hátún et al. 2005; Yashayaev and Seidov 2015) and modeling studies (e.g., Jungclaus et al. 2014) that variability in the amount and temperature of Atlantic water flowing northward across the Greenland–Scotland Ridge are driven, in part, by the strength of the SPG. In the Bergen Climate Model this is resonated in enhanced spectral power at the same frequencies for the modeled heat transport by the NwAC () and the SPG strength (calculated from the barotropic streamfunction in the subpolar region), including the 14-yr periodicity found in ocean heat anomalies within the northern seas (Fig. 9b). A strengthening of the subpolar gyre precedes heat and salt content changes in the Norwegian Sea by 1 year, and , respectively. Our results thus support a close coupling between the subpolar North Atlantic and climate variability in the northern seas through variations in poleward heat transport.

7. Oceanic forcing of atmospheric variability

The identified persistent northward advection of anomalous ocean heat and the consequent decadal changes in surface heat fluxes and surface air temperature (Figs. 6 and 9a) suggest potential climate predictability in the northern seas region. To further demonstrate the predictive capability associated with a variable ocean heat transport, Fig. 10 shows the 2-yr lagged response in northern seas surface heat fluxes (Fig. 10a) and surface air temperature (Fig. 10c) to changes in the Atlantic inflow based on linear regression. Inflow-driven heat fluxes of 5–20 W m−2 (per standard deviation of heat transport) occur offshore of the Norwegian coast and farther north in the marginal ice zone around Svalbard and in the Barents Sea, reflecting fluctuations in the zonal and meridional extent of the Atlantic domain, and hence sea ice extent (Fig. 10b), respectively. The air temperature response is, on the other hand, pronounced over large parts of the northern seas. The magnitude of the atmospheric response to ocean heat anomalies varies from >0.5°C in the marginal ice zone to 0.1°–0.3°C in the southern Norwegian Sea (Fig. 10c), values being similar to temperature anomalies in the ocean (Fig. 4). There is also a significant atmospheric response over land (0.1°–0.4°C), in agreement with Norwegian climate (air temperature) reflecting decadal temperature variability in the Norwegian Sea (e.g., Eldevik et al. 2014), and, more broadly, European continental climate reflecting North Atlantic SSTs (e.g., Sutton and Hodson 2005).

Fig. 10.

The influence of ocean heat transport into the Nordic seas (; black circles) on (a) surface heat flux to the atmosphere, (b) sea ice concentration loss, and (c) surface air temperature, based on linear regression analysis on low-pass-filtered time series. The ocean heat transport leads by 2 yr. Only regressions significant at the 95% confidence level are plotted. Units are given in W m−2, %, and °C, respectively, per std().

Fig. 10.

The influence of ocean heat transport into the Nordic seas (; black circles) on (a) surface heat flux to the atmosphere, (b) sea ice concentration loss, and (c) surface air temperature, based on linear regression analysis on low-pass-filtered time series. The ocean heat transport leads by 2 yr. Only regressions significant at the 95% confidence level are plotted. Units are given in W m−2, %, and °C, respectively, per std().

Understanding the interaction between the ocean and the atmosphere is a prerequisite for understanding and predicting climate variability. To what extent anomalous ocean heat leads to atmospheric circulation changes is not addressed here. Modeling studies have generally considered that the amplitude of the atmospheric response to extratropical large-scale SST anomalies is modest compared with internal atmospheric variability (Kushnir et al. 2002), but this is still a matter of debate and might be both model (Omrani et al. 2014; Smirnov et al. 2015) and time scale dependent (Gulev et al. 2013; Sheldon and Czaja 2014). Ocean heat anomalies in the northern seas can in any case yield a significant regional atmospheric response. In line with our results van der Swaluw et al. (2007), also using data from a preindustrial control run, showed that anomalous heat transport by the NwAC forces the atmosphere by increased heat fluxes in the marginal ice zone (>70°N; cf. our Fig. 10b). The atmospheric response to increased heat transport was associated with a cyclonic pressure anomaly and decreased atmospheric heat transport by baroclinic eddies as a result of a decreased poleward temperature gradient in the atmosphere. Similar mechanisms and impact were also identified by Schlichtholz (2013) based on observational data from the northern seas. The regionally confined anomalous atmospheric circulation in response to decadal-scale ocean-driven sea ice variability in the northern seas, and in particular in the Barents Sea, can also drive larger-scale surface climate variability (e.g., Semenov et al. 2010; Liptak and Strong 2014).

In support of decadal ice–ocean interaction in the Barents Sea, the modeled winter (December–April) sea ice cover in the Barents Sea (70°–81°N, 15°–60°E) has a similar spectrum to that of climate variability in the northern seas (Figs. 5 and 9b), including a dominant time scale of 14 years (not shown). The correlation between the low-pass-filtered heat transport into the Barents Sea () and sea ice extent in the Barents Sea is −0.77, with a time lag of 1–2 years in agreement with observations (Årthun et al. 2012; Onarheim et al. 2015). The decadal-scale oscillation furthermore agrees with Venegas and Mysak (2000) and Vinje (2001), who found fluctuations in the observed Barents Sea ice extent with a time scale of 16–20 and 12–14 years, respectively, related to a variable NwAC.

Identifying a time scale associated with northward propagating ocean heat anomalies from the subpolar North Atlantic toward the Arctic (Figs. 5 and 9b) and their consequent interaction with the atmosphere (Figs. 9a and 10) is essential in terms of potential climate prediction. A recent model study by Escudier et al. (2013) found that a 20-yr coupled mode of atmosphere–ice–ocean variability may exist in the subpolar North Atlantic in which the propagation of thermohaline anomalies from the subpolar gyre interacts with the atmosphere in the northern seas to eventually produce anomalies of the opposite sign in the Labrador Sea. Results demonstrated herein further supports a coupled mode of variability in the northern seas associated with the propagation and atmospheric imprint of ocean heat anomalies. The different time scale in the two models is likely related to the stronger northward heat transport in the Bergen Climate Model (Langehaug et al. 2012b) as the time scale of the cycle is set by the propagation of anomalies along the rim of the Nordic seas. The robustness of the time scale identified herein needs to be further assessed, but we reiterate that the characteristic 14-yr time scale of climate variability in the northern seas is supported by observed SST fluctuations in the Norwegian Sea (Fig. 5c).

8. Conclusions

Interannual- to decadal-scale ocean heat anomalies associated with the northern limb of the Atlantic meridional overturning circulation propagate persistently toward the Arctic (e.g., Holliday et al. 2008). This poleward propagation of anomalous ocean heat is commonly understood to be a primary source for climate predictability (e.g., Latif and Keenlyside 2011). Here, we have used a 500-yr control simulation from the fully coupled Bergen Climate Model (BCM), aided by observed sea surface temperatures (HadISST), to assess the propagation and drivers of ocean heat anomalies in the northern seas (northern North Atlantic, Nordic seas, and Arctic Ocean), and to what extent these anomalies imprint on the atmosphere.

Ocean heat anomalies are found to propagate from the eastern subpolar North Atlantic into and along the rim of the Nordic seas with a speed of 2 cm s−1 (Fig. 3). The characteristic time scale of variability is 14 years, which is also that of observed sea surface temperature variability in the Norwegian Sea during the last century (Fig. 5). The relative roles of ocean and atmosphere in driving ocean heat anomalies are assessed by constructing a depth integrated heat budget for an area covering the Atlantic domain of the Nordic seas (i.e., the Norwegian Sea). Changes in heat content are found to be caused mainly by anomalous ocean heat transport convergence (Fig. 6). Variations in ocean heat convergence largely originate in the inflow from the Atlantic proper, and a temporal decomposition of the Atlantic heat transport shows that volume transport anomalies dominate (Fig. 8). Simulated ocean heat anomalies in the northern seas are thus driven mainly by changes in the strength of the northward flowing Atlantic water. A similar decadal-scale oscillation in the strength of the subpolar gyre (Fig. 9b) further supports the close coupling, observed and modeled, between the subpolar North Atlantic and Nordic seas/Arctic Ocean (e.g., Hátún et al. 2005; Glessmer et al. 2014; Jungclaus et al. 2014).

A potentially predictable relation between anomalous ocean heat and climate in the northern seas region is furthermore identified. Ocean heat anomalies in the northern seas are reflected in regional sea ice extent and found to influence the atmosphere by driving changes in surface air temperatures through anomalous air–sea fluxes (Fig. 10). The interaction with the atmosphere is also most pronounced on a 14-yr time scale. The identified time scale of climate variability, manifested both in anomalous ocean heat transport and its consequent atmospheric response, provides encouraging evidence for climate predictability rooted in the northern seas.

Acknowledgments

This research was supported by the Centre for Climate Dynamics at the Bjerknes Centre for Climate Research through the project PRACTICE and the Research Council of Norway projects EPOCASA and NORTH. We thank Odd Helge Otterå for providing the model data and Helene R. Langehaug for providing the SPG index. We also thank Tore Furevik and three anonymous reviewers for valuable comments that improved the manuscript.

REFERENCES

REFERENCES
Årthun
,
M.
,
T.
Eldevik
,
L. H.
Smedsrud
,
Ø.
Skagseth
, and
R. B.
Ingvaldsen
,
2012
:
Quantifying the influence of Atlantic heat on Barents Sea ice variability and retreat
.
J. Climate
,
25
,
4736
4743
, doi:.
Bleck
,
R.
,
C.
Rooth
,
D. M.
Hu
, and
L. T.
Smith
,
1992
:
Salinity-driven thermocline transients in a wind- and thermohaline-forced isopycnic coordinate model of the North Atlantic
.
J. Phys. Oceanogr.
,
22
,
1486
1505
, doi:.
Bourke
,
R. H.
,
A. M.
Weigel
, and
R. G.
Paquette
,
1988
:
The westward turning branch of the West Spitsbergen Current
.
J. Geophys. Res.
,
93
,
14 065
14 077
, doi:.
Buckley
,
M. W.
,
R. M.
Ponte
,
G.
Forget
, and
P.
Heimbach
,
2014
:
Low-frequency SST and upper-ocean heat content variability in the North Atlantic
.
J. Climate
,
27
,
4996
5018
, doi:.
Burkholder
,
K. C.
, and
M. S.
Lozier
,
2014
:
Tracing the pathways of the upper limb of the North Atlantic meridional overturning circulation
.
Geophys. Res. Lett.
,
41
,
4254
4260
, doi:.
Carmack
,
E.
, and Coauthors
,
2015
:
Toward quantifying the increasing role of oceanic heat in sea ice loss in the new Arctic
.
Bull. Amer. Meteor. Soc.
,
96
,
2079
2105
, doi:.
Carton
,
J. A.
,
G. A.
Chepurin
,
J.
Reagan
, and
S.
Häkkinen
,
2011
: Interannual to decadal variability of Atlantic Water in the Nordic and adjacent seas. J. Geophys. Res., 116, C11035, doi:.
Chelton
,
D. B.
,
1983
: Effects of sampling errors in statistical estimation. Deep-Sea Res., 30, 1083–1103, doi:.
Chepurin
,
G. A.
, and
J. A.
Carton
,
2012
: Subarctic and Arctic sea surface temperature and its relation to ocean heat content 1982–2010. J. Geophys. Res., 117, C06019, doi:.
Czaja
,
A.
, and
J.
Marshall
,
2001
:
Observations of atmosphere–ocean coupling in the North Atlantic
.
Quart. J. Roy. Meteor. Soc.
,
127
,
1893
1916
, doi:.
Déqué
,
M.
,
C.
Dreveton
,
A.
Braun
, and
D.
Cariolle
,
1994
:
The ARPEGE/IFS atmosphere model—A contribution to the French community climate modeling
.
Climate Dyn.
,
10
,
249
266
, doi:.
Deser
,
C.
, and
M. L.
Blackmon
,
1993
:
Surface climate variations over the North Atlantic Ocean during winter—1900–1989
.
J. Climate
,
6
,
1743
1753
, doi:.
Dickson
,
R. R.
,
J.
Meincke
,
S.-A.
Malmberg
, and
A. J.
Lee
,
1988
:
The “great salinity anomaly” in the northern North Atlantic 1968–1982
.
Prog. Oceanogr.
,
20
,
103
151
, doi:.
Drinkwater
,
K. F.
,
M.
Miles
,
I.
Medhaug
,
O. H.
Otterå
,
T.
Kristiansen
,
S.
Sundby
, and
Y.
Gao
,
2014
:
The Atlantic multidecadal oscillation: Its manifestations and impacts with special emphasis on the Atlantic region north of 60°N
.
J. Mar. Syst.
,
133
,
117
130
, doi:.
Eden
,
C.
, and
T.
Jung
,
2001
:
North Atlantic interdecadal variability: Oceanic response to the North Atlantic Oscillation (1865–1997)
.
J. Climate
,
14
,
676
691
, doi:.
Eldevik
,
T.
,
J. E. Ø.
Nilsen
,
D.
Iovino
,
K. A.
Olsson
,
A. B.
Sandø
, and
H.
Drange
,
2009
:
Observed sources and variability of Nordic seas overflow
.
Nat. Geosci.
,
2
,
406
410
, doi:.
Eldevik
,
T.
, and Coauthors
,
2014
:
A brief history of climate—The northern seas from the Last Glacial Maximum to global warming
.
Quat. Sci. Rev.
, 106, 225–246, doi:.
Escudier
,
R.
,
J.
Mignot
, and
D.
Swingedouw
,
2013
:
A 20-year coupled ocean–sea ice–atmosphere variability mode in the North Atlantic in an AOGCM
.
Climate Dyn.
,
40
,
619
636
, doi:.
Farneti
,
R.
, and
G. K.
Vallis
,
2011
:
Mechanisms of interdecadal climate variability and the role of ocean–atmosphere coupling
.
Climate Dyn.
,
36
,
289
308
, doi:.
Francis
,
J. A.
, and
E.
Hunter
,
2007
: Drivers of declining sea ice in the Arctic winter: A tale of two seas. Geophys. Res. Lett., 34, L17503, doi:.
Frankcombe
,
L. M.
,
A.
von der Heydt
, and
H. A.
Dijkstra
,
2010
:
North Atlantic multidecadal climate variability: An investigation of dominant time scales and processes
.
J. Climate
,
23
,
3626
3638
, doi:.
Furevik
,
T.
,
2000
:
On anomalous sea surface temperatures in the Nordic seas
.
J. Climate
,
13
,
1044
1053
, doi:.
Furevik
,
T.
,
2001
:
Annual and interannual variability of Atlantic Water temperatures in the Norwegian and Barents Seas: 1980–1996
.
Deep-Sea Res. I
,
48
,
383
404
, doi:.
Furevik
,
T.
, and
J.
Nilsen
,
2005
: Large-scale atmospheric circulation variability and its impacts on the Nordic Seas ocean climate—A review. The Nordic Seas: An Integrated Perspective, Geophys. Monogr., Vol. 158, Amer. Geophys. Union, 105–136.
Furevik
,
T.
,
M.
Bentsen
,
H.
Drange
,
I. K. T.
Kindem
,
N. G.
Kvamstø
, and
A.
Sorteberg
,
2003
:
Description and evaluation of the Bergen Climate Model: ARPEGE coupled with MICOM
.
Climate Dyn.
,
21
,
27
51
, doi:.
Gaspar
,
P.
,
1988
:
Modeling the seasonal cycle of the upper ocean
.
J. Phys. Oceanogr.
,
18
,
161
180
, doi:.
Glessmer
,
M. S.
,
T.
Eldevik
,
K.
Våge
,
J. E. Ø.
Nilsen
, and
E.
Behrens
,
2014
:
Atlantic origin of observed and modelled freshwater anomalies in the Nordic Seas
.
Nat. Geosci.
,
7
,
801
805
, doi:.
Gulev
,
S. K.
,
M.
Latif
,
N.
Keenlyside
,
W.
Park
, and
K. P.
Koltermann
,
2013
:
North Atlantic Ocean control on surface heat flux on multidecadal timescales
.
Nature
,
499
,
464
467
, doi:.
Hansen
,
D. V.
, and
H. F.
Bezdek
,
1996
:
On the nature of decadal anomalies in North Atlantic sea surface temperature
.
J. Geophys. Res.
,
101
,
8749
8758
, doi:.
Hátún
,
H.
,
A. B.
Sandø
,
H.
Drange
,
B.
Hansen
, and
H.
Valdimarsson
,
2005
:
Influence of the Atlantic subpolar gyre on the thermohaline circulation
.
Science
,
309
,
1841
1844
, doi:.
Hátún
,
H.
, and Coauthors
,
2009
:
Large bio-geographical shifts in the north-eastern Atlantic Ocean: From the subpolar gyre, via plankton, to blue whiting and pilot whales
.
Prog. Oceanogr.
,
80
,
149
162
, doi:.
Helland-Hansen
,
B.
, and
F.
Nansen
,
1909
:
The Norwegian Sea
.
Fiskdir. Skr. Ser. Havunders.
,
11
(
2
),
1
360
.
Holliday
,
N. P.
, and Coauthors
,
2008
:
Reversal of the 1960s to 1990s freshening trend in the northeast North Atlantic and Nordic Seas
.
Geophys. Res. Lett.
,
35
, L03614, doi:.
Horel
,
J. D.
,
1984
:
Complex principal component analysis: Theory and examples
.
J. Climate Appl. Meteor.
,
23
,
1660
1673
, doi:.
Hughes
,
S. L.
,
N. P.
Holliday
,
E.
Colbourne
,
V.
Ozhigin
,
H.
Valdimarsson
,
S.
Østerhus
, and
K.
Wiltshire
,
2009
:
Comparison of in situ time-series of temperature with gridded sea surface temperature datasets in the North Atlantic
.
ICES J. Mar. Sci.
,
66
, 1467–1479, doi:.
Hurrell
,
J. W.
,
1995
:
Decadal trends in the North Atlantic Oscillation: Regional temperatures and precipitation
.
Science
,
269
,
676
679
, doi:.
Jungclaus
,
J. H.
,
K.
Lohmann
, and
D.
Zanchettin
,
2014
:
Enhanced 20th-century heat transfer to the Arctic simulated in the context of climate variations over the last millennium
.
Climate Past
,
10
,
2201
2213
, doi:.
Krahmann
,
G.
,
M.
Visbeck
, and
G.
Reverdin
,
2001
:
Formation and propagation of temperature anomalies along the North Atlantic Current
.
J. Phys. Oceanogr.
,
31
,
1287
1303
, doi:.
Kushnir
,
Y.
,
1994
:
Interdecadal variations in North Atlantic sea surface temperature and associated atmospheric conditions
.
J. Climate
,
7
,
141
157
, doi:.
Kushnir
,
Y.
,
W.
Robinson
,
I.
Bladé
,
N.
Hall
,
S.
Peng
, and
R.
Sutton
,
2002
:
Atmospheric GCM response to extratropical SST anomalies: Synthesis and evaluation
.
J. Climate
,
15
,
2233
2256
, doi:.
Langehaug
,
H. R.
,
I.
Medhaug
,
T.
Eldevik
, and
O. H.
Otterå
,
2012a
:
Arctic/Atlantic exchanges via the subpolar gyre
.
J. Climate
,
25
,
2421
2439
, doi:.
Langehaug
,
H. R.
,
P. B.
Rhines
,
T.
Eldevik
,
J.
Mignot
, and
K.
Lohmann
,
2012b
: Water mass transformation and the North Atlantic Current in three multicentury climate model simulations. J. Geophys. Res., 117, C11001, doi:.
Latif
,
M.
, and
N. S.
Keenlyside
,
2011
:
A perspective on decadal climate variability and predictability
.
Deep-Sea Res. II
,
58
,
1880
1894
, doi:.
Liptak
,
J.
, and
C.
Strong
,
2014
:
The winter atmospheric response to sea ice anomalies in the Barents Sea
.
J. Climate
,
27
,
914
924
, doi:.
Liu
,
Z.
,
2012
:
Dynamics of interdecadal climate variability: A historical perspective
.
J. Climate
,
25
,
1963
1995
, doi:.
Lohmann
,
K.
,
H.
Drange
, and
M.
Bentsen
,
2009
:
Response of the North Atlantic subpolar gyre to persistent North Atlantic oscillation like forcing
.
Climate Dyn.
,
32
,
273
285
, doi:.
Lohmann
,
K.
, and Coauthors
,
2014
:
The role of subpolar deep water formation and Nordic Seas overflows in simulated multidecadal variability of the Atlantic meridional overturning circulation
.
Ocean Sci.
,
10
,
227
241
, doi:.
Mauritzen
,
C.
,
1996
:
Production of dense overflow waters feeding the North Atlantic across the Greenland-Scotland Ridge. Part 1: Evidence for a revised circulation scheme
.
Deep-Sea Res. I
,
43
,
769
806
, doi:.
Medhaug
,
I.
,
H. R.
Langehaug
,
T.
Eldevik
,
T.
Furevik
, and
M.
Bentsen
,
2012
:
Mechanisms for decadal scale variability in a simulated Atlantic meridional overturning circulation
.
Climate Dyn.
,
39
,
77
93
, doi:.
Meehl
,
G. A.
, and Coauthors
,
2014
:
Decadal climate prediction: An update from the trenches
.
Bull. Amer. Meteor. Soc.
,
95
,
243
267
, doi:.
Mork
,
K. A.
,
Ø.
Skagseth
,
V.
Ivshin
,
V.
Ozhigin
,
S. L.
Hughes
, and
H.
Valdimarsson
,
2014
:
Advective and atmospheric forced changes in heat and fresh water content in the Norwegian Sea, 1951–2010
.
Geophys. Res. Lett.
, 41, 6221–6228, doi:.
Nilsen
,
J. E. Ø.
, and
E.
Falck
,
2006
:
Variations of mixed layer properties in the Norwegian Sea for the period 1948–1999
.
Prog. Oceanogr.
,
70
,
58
90
, doi:.
Omrani
,
N.-E.
,
N. S.
Keenlyside
,
J.
Bader
, and
E.
Manzini
,
2014
:
Stratosphere key for wintertime atmospheric response to warm Atlantic decadal conditions
.
Climate Dyn.
,
42
,
649
663
, doi:.
Onarheim
,
I. H.
,
L. H.
Smedsrud
,
R. B.
Ingvaldsen
, and
F.
Nilsen
,
2014
:
Loss of sea ice during winter north of Svalbard
.
Tellus
,
66A
, 23933, http://dx.doi.org/10.3402/tellusa.v66.23933.
Onarheim
,
I. H.
,
T.
Eldevik
,
M.
Årthun
,
R. B.
Ingvaldsen
, and
L. H.
Smedsrud
,
2015
:
Skillful prediction of Barents Sea ice cover
.
Geophys. Res. Lett.
,
42
,
5364
5371
, doi:.
Orvik
,
K. A.
,
Ø.
Skagseth
, and
M.
Mork
,
2001
:
Atlantic inflow to the Nordic Seas: Current structure and volume fluxes from moored current meters, VM-ADCP and SeaSoar-CTD observations, 1995–1999
.
Deep-Sea Res. I
,
48
,
937
957
, doi:.
Otterå
,
O. H.
,
M.
Bentsen
,
I.
Bethke
, and
N. G.
Kvamstø
,
2009
:
Simulated pre-industrial climate in Bergen Climate Model (version 2): Model description and large-scale circulation features
.
Geosci. Model Dev.
,
2
,
197
212
, doi:.
Otterå
,
O. H.
,
M.
Bentsen
,
H.
Drange
, and
L.
Suo
,
2010
:
External forcing as a metronome for Atlantic multidecadal variability
.
Nat. Geosci.
,
3
,
688
694
, doi:.
Polyakov
,
I. V.
, and Coauthors
,
2005
:
One more step toward a warmer Arctic
.
Geophys. Res. Lett.
,
32
, L17605, doi:.
Rayner
,
N. A.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
,
4407
, doi:.
Rhines
,
P.
,
S.
Häkkinen
, and
S.
Josey
,
2008
: Is oceanic heat transport significant in the climate system? Arctic–Subarctic Ocean Fluxes, R. R. Dickson, J. Meincke, and P. Rhines, Eds., Springer, 87–109.
Rudels
,
B.
,
H. J.
Friedrich
, and
D.
Quadfasel
,
1999
:
The Arctic circumpolar boundary current
.
Deep-Sea Res. II
,
46
,
1023
1062
, doi:.
Salas-Mélia
,
D.
,
2002
:
A global coupled sea ice–ocean model
.
Ocean Modell.
,
4
,
137
172
, doi:.
Sandø
,
A.
,
J.
Nilsen
,
T.
Eldevik
, and
M.
Bentsen
,
2012
:
Mechanisms for variable North Atlantic–Nordic seas exchanges
.
J. Geophys. Res.
,
117
, C12006, doi:.
Schlichtholz
,
P.
,
2013
:
Observational evidence for oceanic forcing of atmospheric variability in the Nordic seas area
.
J. Climate
,
26
,
2957
2975
, doi:.
Screen
,
J. A.
,
C.
Deser
,
I.
Simmonds
, and
R.
Tomas
,
2014
:
Atmospheric impacts of Arctic sea-ice loss, 1979–2009: Separating forced change from atmospheric internal variability
.
Climate Dyn.
, 43, 333–344, doi:.
Semenov
,
V. A.
,
M.
Latif
,
D.
Dommenget
,
N. S.
Keenlyside
,
A.
Strehz
,
T.
Martin
, and
W.
Park
,
2010
:
The impact of North Atlantic–Arctic multidecadal variability on Northern Hemisphere surface air temperature
.
J. Climate
,
23
,
5668
5677
, doi:.
Sheldon
,
L.
, and
A.
Czaja
,
2014
:
Seasonal and interannual variability of an index of deep atmospheric convection over western boundary currents
.
Quart. J. Roy. Meteor. Soc.
,
140
,
22
30
, doi:.
Skagseth
,
Ø.
, and
K. A.
Mork
,
2012
:
Heat content in the Norwegian Sea, 1995–2010
.
ICES J. Mar. Sci.
,
69
,
826
832
, doi:.
Smedsrud
,
L. H.
, and Coauthors
,
2013
:
The role of the Barents Sea in the Arctic climate system
.
Rev. Geophys.
,
51
,
415
449
, doi:.
Smirnov
,
D.
,
M.
Newman
,
M. A.
Alexander
,
Y.-O.
Kwon
, and
C.
Frankignoul
,
2015
:
Investigating the local atmospheric response to a realistic shift in the Oyashio sea surface temperature front
.
J. Climate
,
28
,
1126
1147
, doi:.
Sutton
,
R. T.
, and
M. R.
Allen
,
1997
:
Decadal predictability of North Atlantic sea surface temperature and climate
.
Nature
,
388
,
563
567
, doi:.
Sutton
,
R. T.
, and
D. L. R.
Hodson
,
2005
:
Atlantic Ocean forcing of North American and European summer climate
.
Science
,
309
,
115
118
, doi:.
Terray
,
L.
, and
O.
Thual
,
1995
:
Oasis: le couplage océan–atmosphere
.
Meteorologie
,
10
,
50
61
.
van der Swaluw
,
E.
,
S.
Drijfhout
, and
W.
Hazeleger
,
2007
:
Bjerknes compensation at high northern latitudes: The ocean forcing the atmosphere
.
J. Climate
,
20
,
6023
6032
, doi:.
Venegas
,
S. A.
, and
L. A.
Mysak
,
2000
:
Is there a dominant timescale of natural climate variability in the Arctic?
J. Climate
,
13
,
3412
3434
, doi:.
Vihma
,
T.
,
2014
:
Effects of Arctic sea ice decline on weather and climate: A review
.
Surv. Geophys.
, 35, 1175–1214, doi:.
Vinje
,
T.
,
2001
:
Anomalies and trends of sea-ice extent and atmospheric circulation in the Nordic seas during the period 1864–1998
.
J. Climate
,
14
,
255
267
, doi:.
Visbeck
,
M.
,
H.
Cullen
,
G.
Krahmann
, and
N.
Naik
,
1998
:
An ocean model’s response to North Atlantic Oscillation-like wind forcing
.
Geophys. Res. Lett.
,
25
,
4521
4524
, doi:.
Williams
,
R. G.
,
V.
Roussenov
,
D.
Smith
, and
M. S.
Lozier
,
2014
:
Decadal evolution of ocean thermal anomalies in the North Atlantic: The effects of Ekman, overturning, and horizontal transport
.
J. Climate
,
27
,
698
719
, doi:.
Yashayaev
,
I.
, and
D.
Seidov
,
2015
:
The role of the Atlantic Water in multidecadal ocean variability in the Nordic and Barents Seas
.
Prog. Oceanogr.
,
132
,
68
127
, doi:.