Long‐Term Moored Current and Temperature Measurements of the Atlantic Inflow Into the Nordic Seas in the Norwegian Atlantic Current; 1995–2020

Interannual variability of the Atlantic inflow (AI) into the Nordic Seas is studied using moored current and temperature measurements in the Norwegian Atlantic Current during the 25‐year period 1995–2020. We show that the properties of the AI are strongly connected to the dynamics of the Northern Atlantic and demonstrate the robustness of the average winter wind field in driving the variability with 3‐month winter maxima in the range of (40–50) cms−1. In that sense, the zero‐wind‐stress‐curl (ZWSC)‐lines aligned northeastward are crucial. Occasional events with slowdown of winter‐maxima are connected to a meridional and zonal shift of the ZWSC‐lines, and not the strength of the wind. This is manifested in a correlation of 0.3 versus the North Atlantic Oscillation (NAO) winter‐index, demonstrating that the NAO‐index is insufficient to explain the AI‐variability. The records show no long‐term‐trend over the 25‐year period 1995–2020, regarding volume flux, temperature, nor temperature flux.


of 10
Sea with a minor inflow northward through the Denmark Strait. The second branch flows through the Rock all Trough, entering the Nordic Seas through the Faroe Shetland Channel (Chafik et al., 2014;Childers et al., 2015;Fratantoni, 2001;Orvik & Niiler, 2002;Reverdin et al., 2003). The inflow then continues as the two-branch NwAC through the entire Norwegian Sea toward the Arctic Ocean (Orvik & Niiler, 2002;Poulain et al., 1996). The western branch is a jet in the Polar Front of the interior Norwegian Sea whilst the eastern branch-the Norwegian Atlantic Slope Current (NwASC)-is a nearly barotropic shelf edge current along the Norwegian shelf break that tends to flow into the Barents Sea and Arctic Ocean. Accordingly, the NwASC is the major link between the North Atlantic and Arctic Ocean, whilst the western branch appears to feed the interior Nordic Seas and mainly retroflects south as it approaches the Fram Strait (Walczowski & Piechura, 2007).
This study concentrates on the NwASC, the major inflow branch accounting for about 60% of the AI (Orvik et al., 2001). We utilize the moored current meters and temperature measurements in the Svinøy section located Figure 1. Schematic of the major pathways of near-surface Atlantic water in the Northern Atlantic and Nordic Seas toward the Arctic in the context of superimposed sea surface temperature (°C) from an AVHHR image in March. The straight line shows the Svinøy section where the mooring sites are indicated in the Norwegian Atlantic Slope Current. The lower right panel shows the slope area with the initial four moorings and current meters (*) and the long-term mooring S1. 1-year mean current (full lines) and temperature (dashed lines) 1997-1998 are also shown. Abbreviations are explained in the text (after Orvik & Niiler, 2002;Orvik & Skagseth, 2005). about 300 km downstream from the Faroe Shetland Channel. This line cuts through the NwAC at 62°N near its entrance to the Northern areas and captures as such the entire AI into the Nordic Seas ( Figure 1). In this section the two-branch flow is well established as a shelf-edge current and a frontal jet (Orvik et al., 2001), and is thus a suitable place for monitoring the AI.
A series of studies have been carried out over the past decades to investigate the structure, transport, and longterm variability of this crucial AI, both modeling work and studies based on a variety of observations (Bringedal et al., 2018;Chafik et al., 2014;Mork & Skagseth, 2010;Olsen et al., 2008;Orvik et al., 2001;Østerhus et al., 2019;Rossby et al., 2018;Sandø et al., 2012). There has been nearly a continuous effort to gain quantitative assessment of estimated mean volume and temperature fluxes. Recent estimates of the average AI volume flux are 8.0 Sv (Østerhus et al., 2019) and 7.7 Sv (Rossby et al., 2018), while a former estimate in the Svinøy section was 7.6 Sv (Orvik et al., 2001; Sv = 10 6 m 3 s −1 ). However, long-term studies are commonly based on numerical models or partly on combinations of in situ measurements and proxies as altimetry data (Berx et al., 2013). To our knowledge, the 25-year continuous time series in the Svinøy section, is the only long-term record based solely on moored current and temperature measurements. We take advantage of these hourly measurements in the core of the NwASC to investigate variabilities on interannual timescales. We consider the climatological wind field over the subpolar North Atlantic (SPNA) as a major driver of the AI and assess how its anomalies-the North Atlantic Oscillation (NAO) and East Atlantic pattern (EAP)-determine its variability. Temperature and temperature fluxes, and their link to large-scale anomalies of the SPNA, are also considered. They are crucial both with respect to climate change and ecology in Northern areas.
The study is an extension of Orvik and Skagseth (2005) and is organized as follows. In Section 2 the data basis and methodology applied are presented. Then the major findings are presented in Section 3. They are put in perspective in the discussion in Section 4 by connecting the variability of the AI to upstream forcing mechanisms, dynamics, and water mass properties. Section 5 closes with concluding remarks.

Data and Methodology
We apply the same methodology as in Orvik and Skagseth (2005) with reference therein for further details, to utilize the moored measurements; located in the core of the NwASC in the Svinøy section at position 62°48′N, 4°55′E ( Figure 1). We use hourly current and temperature records from "one single" Aanderaa RCM-7 current meter at 100 m depth with a data recovery close to 100% for the 25-year period April 1995 to May 2020. Concerning volume flux of AW in the NwASC, use of the 100 m current record is justifiable because the volume flux (V in Sv) calculated from an array of current meters can be estimated from a single current meter (v in ms −1 ) at 100 m with a correlation coefficient of 0.86, as V = α10 8 v, where, α = 0.13 m 2 (Orvik & Skagseth, 2003b). To optimize the hourly data for analysis on seasonal and interannual timescales, we apply boxcar moving-average low pass filters with 30 days (1 month), 90-day (3 months), and 365-days (1 year) cut-off periods to the original time series. The heat flux of the NwASC is given by Q = ∫ S c p ρvTds with 0°C as reference temperature, where ρ is density of sea water, c p the specific heat capacity, T the temperature, and v the along-slope current velocity perpendicular to the area S of the flow. Since the volume flux V = ∫ S vds can be determined by using a single current meter record as V = α10 8 v, the oceanic heat flux estimate is referred to as temperature flux Q T = ρc p α10 8 vT.
In considering anomaly variations, we split the total v, T, and vT into 25-year averages and anomaly parts as =̄+ ′, =̄+ ′ , and =̄+ ( )′ , where ( )′ =̄′ + ′̄+ ′ ′ . An overview of the data recovery is dispatched in Figure 2 in terms of hourly current and temperature time series superimposed the 30-day moving averages.

Results
The time series of hourly along-slope currents depict in Figure 2a  The temperature anomaly (T′) in Figure 3b also shows a prominent seasonal cycle with amplitude of nearly 1.0°C about the 25-year mean of 8.9°C, and maximum of 10.2°C in late fall. The coincidence in seasonality of v on T is manifested in a correlation of 0.56, v trailing T with 100 days. On interannual timescale, the most striking feature is the well-known temperature increase of about 1°C over the first 10 years, reaching a maximum in 2003 (Orvik & Skagseth, 2005). After that warming-event the temperature became fairly stable till 2009, when a cooling trend started reaching an absolute 3-month minimum of 7.7°C in 2018, comparable with the 1997 minimum. Superimposed on this cooling trend, is an extraordinary cooling during the period 2014-2018, resulting in a 1°C temperature decrease over 3-4 years. Overall, there is an insignificant trend of 0.0066°C/year (using annual data) over the 25-year period 1995-2020.
The average temperature flux in Figure 3c is 165 TWatt, varying in the range of (130, 203) TWatt on interannual timescale. It exhibits an increasing trend from 1997 until 2006, and then a subsequent decrease until 2018, resulting in no overall trend.

Current Velocity
It is well established that the variability of the AI is mainly driven by the atmospheric forcing over the Northern Atlantic (Bringedal et al., 2018;Orvik & Skagseth, 2003a;Richter et al., 2009;Sandø & Furevik, 2008;Sandø et al., 2012). The average wind field has a prominent seasonality with strong westerly winds toward the northeast during winter and is less dynamic with moderate winds during summer. Here, we concentrate on how variabilities on the average winter wind field (AWWF) over the SPNA, affect the AI (defining AWWF as the -average) by applying NCEP/NCAR sea-level-pressure (SLP)-data (Kalnay et al., 1996). The SLPlines and associated zero-wind-stress-curl (ZWSC)-lines of the AWWF exhibit a pronounced poleward tilt from 6 of 10 SW to NE (Figure 4a). To its north the wind-stress-curl is positive, driving the cyclonic subpolar gyre (SPG), and to its south it is negative, driving the anticyclonic subtropical gyre (STG). The ZWSC-line is crucial in marking the confluence zone of the two gyres. This is because the wind-stress-curl driving the Ekman-pumping, builds up an integrated pressure gradient resulting in a maximum geostrophic flow along ZWSC-lines. Prominent anomalies superimposed on the AWWF are identified as the NAO and EAP. The NAO is a measure of the strength of the westerlies and the EAP the location or orientation of the westerlies-often interpreted as a southward shift of 7 of 10 the NAO pattern (Foukal & Lozier, 2017). The EAP is associated with wind-stress-curl anomalies that induce Ekman divergence in the SPG and convergence in the STG, modulating both gyres in phase. We demonstrate the crucial role of the meridional and zonal shift of the wind field over the SPNA, in determining the variability of the AI on interannual timescale. A case study is performed by comparing the winter maxima/minima of the AI with the strength and shift of the wind field for the "NAO-winter" December-March, selecting absolute winter maxima/minima in 2006-2011/2001-2018 (Figure 4). The associated wind fields in terms of SLP and associated ZWSC-lines are depicted in Figures 4b-4e.
The wind patterns during maximum AI-winters (Figures 4b and 4c) show SLP and ZWSC-lines in a pronounced tilt from SW to NE aligned with the AWWF, resulting in a poleward NAC. For the minimum AI-winters in 2001/2018, there is a shift of the wind field toward zonally running ZWSC-lines in (Figures 4d and 4e), yielding a more zonal NAC (Marshall et al., 2001). Considering the combination of NAO/EAP winter-phases ( 3 for the NAO-index versus v, corroborates that the NAO-index is insufficient to explain the AI-variability, and shows that a SLP index is not always adequate to describe the ocean response to the atmospheric variability. Over the 25 years, the 3-month winter maxima vary mainly in the range of (40-50) cms −1 with only two significant exceptions below the overall average of 35 cms −1 with apparent loss of seasonality (2001/2018). This substantiates the robustness of the climatological pattern and that the poleward alignment of ZWSC-lines being crucial in driving the AI, rather than the strength of the wind (Marshall et al., 2001).
Concerning interannual variability, the current is fairly stable with variabilities of about 10% of the average flow and much less than for the seasonal signal. According to Figure 3, it is evident that the interannual maxima coincide with years of winter maxima, for example, in 1998, 2005-2006, and 2011, and minima when the winter maxima are low for example, in 2001 and 2018. This concurrence substantiates that they are affected by the same driving mechanism. With respect to the long-term trend, the AI shows an increase of 10 cms −1 (10%) toward absolute maximum during 1996-2006 followed by a similar decrease over the next 14 years back to the 1996 level, resulting in an insignificant trend of −0.057 cms −1 /year (from annual data) over the 25-year period 1995-2020. This agrees with the development of the Gulf Stream, showing no long-term trend over the 20-year period 1992-2012 .

Temperature
The properties of the AI are strongly connected to the hydrography of the eastern SPNA, and the mechanisms causing its variability, because the eastern SPNA is the entranceway for sub-tropical water into subpolar latitudes (Foukal & Lozier, 2018). This connection is manifested by the coincidence between the AI temperature record (Figure 3b) and the ocean heat content (OHC) of the eastern SPNA during 1990-2015, showing a strong increase during 1994-2004, then a subsequent cooling during 2005-2015 with intensification from 2013 (Piecuch et al., 2017). Thus, the two most striking events observed in the eastern SPNA are captured in the Svinøy section 1-2 years later. The cooling event in the Svinøy section is consistent with freshening (Orvik at al., 2001), and subsequent freshening of the Nordic Seas since 2011 (Mork et al., 2019).
The extraordinary warming and salinity-increase during 1995-2004 have been studied extensively both in the AI and the SPNA (Bersch et al., 2002;Hatun et al., 2005;Häkkinen et al., 2011). They highlight the dynamics of the SPG in controlling the composition and strength of the AI through shifts in the subpolar front and advective pathways toward the Iceland-Scotland Ridge. Specifically, a weakening and westward contraction of the SPG, opening the gateway for throughput of warm water from the STG. Hatun et al. (2005) for example, showed that the salinity-increase of the AI on the Iceland-Scotland Ridge and thus the warming captured later in the Svinøy section, coincided with a decline of the SPG circulation with a 1-year time lag, using an altimetry-based gyre index (Häkkinen & Rhines, 2004). However, the gyre index continued to decline until 2015, despite a cooling of the eastern SPNA from 2005. Considering that, Foukal and Lozier (2017), (2018) found that the properties of the eastern SPNA do not covary with the SPG size and suggested that the SPG dynamics do not control the strength of the inter-gyre throughput.
The most intriguing development of the AI-temperature is the cooling from an absolute maximum in 2014 to a subsequent minimum in 2018 (Figure 3b). This cooling starts 3-4 years after the extraordinarily freshening event in the Newfoundland Basin during 2012-2016, explained as an outbreak of Arctic water from the Labrador Current into the Newfoundland Basin (Holliday et al., 2020). The anomaly joined the NAC and propagated eastward and subsequently northward being identified as a salinity/temperature anomaly in the Iceland Basin during 2014-2018 (Reverdin et al., 2018), before entering the Nordic Seas. However, this interpretation has been questioned by Kenigson and Timmermans (2021), arguing that also this freshening can be explained by the SPG dynamics. The conflicting views of the mechanism behind the water-properties of the eastern SPNA and thus the AI, demonstrates that the complexity of the underlying mechanisms is not fully resolved.

Temperature Flux
The variability of the temperature flux ( Figure 3c) shows a remarkable coincidence with the volume flux (V). This is manifested by a correlation of 0.96 versus V and 0.32 versus T, respectively. The dominating V-effect is demonstrated in Figure 3c, where time series of each anomaly term are presented in conjunction with vT (Orvik & Skagseth, 2005). T shows a modulating effect on longer timescales, particularly during the extraordinary warming 1996-2004 and contrasting cooling 2014-2018. These findings agree with Foukal and Lozier (2018), showing that the variability of the subtropical-origin temperature flux into the eastern SPNA is primarily driven by transport, with a secondary role from temperature.
On interannual time scales, the temperature flux increased from 1996 to 2006, then decreased from 2006 to 2018, such that the 2018 flux was the same as 1996. Overall, there is no long-term trend over the 25-year period 1995-2020. These findings contradict the common understanding of increasing ocean heat flux into the Nordic Seas, for example, shown by Tsubouchi et al. (2020). However, it is consistent with Mork et al. (2019), who found that the relatively high OHC in the Norwegian Sea despite reduced temperature of the AI, was caused by reduced local air-sea heat fluxes. Overall, the temperature-and volume flux time series in Figure 3 show a concurrent pattern with an increase over the first decade and then a subsequent decrease at the end of the period, after a fairly stable intermediate period.

Concluding Remark
We have shown the core of the NwASC has no long-term trend of AI during the 25 years from 1995 to 2020, regarding volume flux, temperature, nor temperature flux. The variability of the temperature flux is mainly determined by the volume flux with temperature as a secondary effect. We consider the drivers of the AI-variability in light of a conceptual approach by Marshall et al. (2001) and demonstrate the robustness of the AWWF in driving the interannual variability. We found the alignment of the ZWSC-lines to be crucial, and not the strength of the wind field, according to Marshall et al. (2001). Occasional changes of the AI appear to be driven by meridional and zonal shifts of the ZWSC-lines, associated with the EAP. However, in a more complete study we must take the complex topography of the SPNA into account, regarding topographic steered currents and variable wind fields. Effects amplifying the AI as piling up of water toward the British Isles driven by Ekman transport or topographic Sverdrup transport, have to be considered (Bringedal et al., 2018;Skagseth, 2004). An apparent limitation of this study is the emphasis of the major AI into the Nordic Seas-the NwASC, and not including the western branch (Orvik et al., 2001). Since the western branch is established as a jet in the Iceland Faroe Front continuing into the Nordic Seas and partly joining the NwASC, it appears to be connected to the same upstream conditions as the NwASC. Based on our knowledge to date, there is no reason to anticipate different variability than shown in Orvik et al. (2001). This is still an open question, so further studies are needed.

Data Availability Statement
The current and temperature data used are available from the Norwegian Marine Data Centre (NMDC; at http://metadata.nmdc.no/UserInterface/#/). This study is a contribution to the Svinøy section monitoring program, initiated in 1995 and is still in progress. The program was funded initially by the Norwegian Research Council (NFR), and a series of EU-and NFR-programs have contributed. Support from the Norwegian Deepwater Program has been invaluable in continuation of the program; thanks are due to Hans Jørgen Saetre. Also, thanks to the crew onboard R/V Håkon Mosby, and K. Bonnevie for their seamanship on about 70 service cruises. Finally, thanks to Steinar Myking for taking care of the instrumentation and Øystein Skagseth for the data analysis and participation in numerous discussions.