the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 26 Jun 2025
Reduced Cooling in the Norwegian Atlantic Slope Current: Investigating mechanisms of change from 30 years of observations
Abstract. The Norwegian Atlantic Current (NwAC) is a principal conduit for poleward heat and salt transport within the Atlantic Meridional Overturning Circulation (AMOC) and plays a key role of water mass transformation in the Nordic Seas. Its variability exerts a critical influence on high-latitude climate, Arctic Ocean inflows, and deep-water formation in the Nordic Seas. This study presents a comprehensive analysis of a 30-year (1993–2022) hydrographic dataset from four repeat sections across the NwAC, spanning from the southern Norwegian Sea (62.8° N) to Bjørnøya (74.5° N). Hydrographic measurements of temperature and salinity, along with derived relative geostrophic velocities, were combined with surface geostrophic currents from satellite altimetry to obtain absolute geostrophic velocities throughout the water column at each section. This allows us to robustly define the current core of the NwAC and assess its properties. The data reveal substantial variability in water properties and transport across seasonal to multi-annual timescales, alongside significant warming trends. While the cooling and freshening of Atlantic Water (AW) along the Norwegian coast is a persistent feature, our analysis indicates a decreasing cooling trend north of Lofoten (69° N). We examine three potential drivers of this reduced cooling: (1) increased advection speed within the current core, (2) reduced lateral heat loss due to decreasing eddy-activity, and (3) decreased air-sea heat fluxes. We find no evidence for any changes in eddy kinetic energy, but both increased advection speed and reduced air-sea heat loss may contribute to the observed decline in cooling. Simple box model estimates suggest that while neither of the two factors can explain all variability observed in the cooling north of Lofoten, changed heat fluxes can quantitatively account for the long term trends. Our results imply a northward amplification of AW warming along the northern rim of the Atlantic Overturning Circulation
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(1921 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1921 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-2854', Anonymous Referee #1, 19 Aug 2025
-
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2854/egusphere-2025-2854-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
-
RC2: 'Comment on egusphere-2025-2854', Anonymous Referee #2, 03 Sep 2025
The paper is interesting, carrying an analysis of a massive series of repeated sections across the Norwegian Atlantic sloope Current, and the results seem to have been thoroughly investigated and supported.
I nonetheless have a series of comments and questions, that may be helpful to better grasp the main results.:
In AW current core for time series the top 50 m are removed (and thus also Ekman…), as well as constrain on maximum bathymetric depth (for SI and GI), in addition to constrain on S 35.00 (for practical salinity; for SA 35.16). What justifies not taking the top 50 meters? (or did I mis-understood?)
Then, the authors separate the core and non-core parts of this transport (based either on largest current or largest S, depending of the section to 50% of the transport…). I am not exactly sure of the practical interest to make this separation (except after to show that the results are someway independent on the part taken?). Upstream AW properties change, so what? Maybe it could be more straightforward to show the core+non-core, and then say (or in an appendix) that results are results hold in core (and non-core) that for the whole core+non-core (is that correct? I don’t think that this is shown or really discussed; indeed, in the abstract, it seems that there is more emphasis on what happens in the core, but after in the presentation, there is often a symmetry of results presented core and non-core)
Afterwards, even when describing the average core and non-core properties, one sees how each criteria (currents, bathymetry, salinity) strongly constrains the spatial extent of core and non-core; It seems to me that this might complicate a bit complicated the comparison of variations in different sections (in particular, between Svinoy and Gimsoy).
The main result (table 1 and figure 5) is that there is a decrease in the cooling between the sections off Norway and the northern sections, in addition to the multi-decadal variability found in all the time series (more for S than T, although in T I see it more in the non-core time series than in the core time series). The authors find reduced heat loss (sensible heat loss decreasing) and faster advection speed.
Question: what are uncertainties in the heat flux products and to which degree this can be taken at face value (maybe because it is on sensible component, and TA is better and more consistently reproduced in the reanalyses). I am not 100% sure that I fully understand the second result, which is not that clear to me in the time series from geostrophic velocity (and estimate of eddy transfer). Indeed, this interesting result is based not on the currents but on the lag correlations for salinity.
On Fig. 7, I did not fully understand how the vertical shear in the core is estimated and how it is relevant. In some places, cores defined by velocity, so it is important to specify in which layer this vertical shear is estimated: is the layer constant in time or varying? Is it also averaged across the core? Depending on how it is defined, the interpretation of the change is different (negative shear also means higher northward velocity at depth: is that correct?)
For interannual variability, figure 8 seems rather convincing (except for a little less than 10 years out of 31), with, if I understand correctly, advection maintained constant? I am no sure if trend removed? (I don’t think so, but that was not so clear to me)
Minor comments
- 34 Bifurcate
- 40: effect? Probably ‘result in’
- 100: if I understand the approach correctly, smoothing just on depths of isopycnals on individual section (not isopycnal T and S). What is done close to the sea surface? (or when an isopycnal outcrops?)? or near the boundaries (when isopycnal section crosses bottom…)
I understand what is done near bottom for extending T and S, and computing then Relative G V, but how is near-bottom transport then estimated? (I suspect taking exact bathymetry and multiplying by an area?)
And then, step of using at core value (0-error) the altimetry-derived velocity (l.113) to have absolute GV. This seems a reasonable approach. However, what are the errors (resolution; spatial smoothing of altimetric product? Is it similar to the one done on the isopycnal depths? )
- 204-209, I am wondering whether it would be better to present the trends difference Bjornoya-Svinoy an BSO-Svinoy, and not the other way around. Well, both can be argued… (saying than north warms more than south, thus less heat loss or other heat mechanism, seems to me a little bit easier to grasp, as water flows from south to north).
Heat fluxes, first l. 150, but after when the plots are shown, I would adopt the ocean convention as heat flux towards the ocean. Here seems that it is heat flux towards the atmosphere (as seen in figures). Thus, with the convention adopted here, an ocean heat loss is associated with a positive heat flux. I prefer the opposite, as is the vast majority of oceanographic papers that I looked at to verify.
- 258: ‘(Figure 7a,b)
Citation: https://doi.org/10.5194/egusphere-2025-2854-RC2 -
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2854/egusphere-2025-2854-AC1-supplement.pdf
-
EC1: 'Comment on egusphere-2025-2854', Meric Srokosz, 15 Sep 2025
If the reviewers' comments are addressed fully and convincingly in a revised manuscript this would become a publishable paper.
Citation: https://doi.org/10.5194/egusphere-2025-2854-EC1 -
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2854/egusphere-2025-2854-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-2854', Anonymous Referee #1, 19 Aug 2025
Review of Baumann et al.
This work gives valuable updates on transports and hydrographic variability, including long-term trends, at standard monitoring sections along the Norwegian slope. Focus is on reduced cooling at the northern part of this slope, which is most directly linked to declining air-sea heat fluxes. The manuscript is generally well written although too much detail is provided in places. Weaker/less robust aspects of this work (eddy fluxes, advection times) occupy too much space, and the apparent ignorance of the outer Norwegian Atlantic Front Current may be a severe limitation.
If the less robust parts are trimmed and the aforementioned severe limitation addressed (as suggested below), this could become a strong contribution to the literature.
Surface heat flux analysis
Throughout this manuscript, it is stated that the reduced cooling is due to reduced surface heat fluxes (especially sensible heat fluxes). This is introduced as early as line 50 (although disputed already in line 58). I suggest that this main result should be better emphasized, while that the weaker components (time lag and lateral eddy fluxes) should be downplayed.
I would structure the Results as follows:
Section 3.1 is well-executed, and should remain first. This section is, however, somewhat lengthy; some details could/should be moved to the Data & Methods part. The associated figure 3 and 4 are appropriate.
Section 3.2 makes an important contribution.
Section 3.3 discusses trends in temperature differences between the standard sections, relying only on references to Table 1, with no figure. The temporal variability in the temperature differences between the Gimsøy Section (GI) and the Barents Sea Opening (BSO) is not substantively presented in Section 3.4 and Figure 8.
Suggestion: Integrate the information in Section 3.4 into 3.3, so that Table 1 and Figure 8 provide mutual support.
Section 3.3.1 describes the most plausible driver; therefore give this sufficient emphasis.
Section 3.3.2 presents a rather heuristic attempt to discuss eddy fluxes. This is based on current meter moorings at the Svinøy Section, which is far upstream from the main signal (the reduced lateral cooling north of Gimsøy).
Suggestion: Move this section to your Discussion rather than presenting it as a primary result.
Section 3.3.3 is excessively long and detailed, without providing commensurate value to the paper.
Suggestion: Consider retaining this section in abbreviated form, showing only Figure 7b for the GI–BSO region, where your main signal is identified (e.g. Figure 8). The connections between the advective time lags and the discrepancy between the GI-BSO temperature difference and sea-to-air heat fluxes are not particularly strong. However, it is worth showing that the lags were short immediately after 2000 and after 2017, when the aforementioned discrepancy was large.
Inner and outer current branches
It is somewhat surprising that this work does not refer to Blindheim et al. (2000), whose abstract, amongst other findings, states: “A temperature rise in the narrowing Norwegian Atlantic Current is strongest in the north.”
This ‘narrowing’ refers to the outer branch (NwAFC), which is not addressed in the present work. This represents a potentially serious limitation. The spatial windows for estimating time series at the GI and SI sections represent a predominantly barotropic core/slope current. This is, however, not the case for the BSO, where the window is much broader, likely including influence from the outer branch, and represents much less of a (likely swift) slope current. Parts of the NwAFC passing through the Svinøy section, will turn eastwards with the topography of the Vøring Plateau, feeding the outer branch at GI; and likely also contributing to parts of the area selected for the BSO. The reported advection time lags of up to 12 months between the relatively closely positioned GI and BSO likely involve this outer current branch, rather than solely a fast nearly barotropic slope current, as portrayed here. This alternative (and in my opinion more realistic) perspective including the outer branch would require adjustments to your box model. Most importantly would the surface area of the Atlantic waters (Bsurf) become variable, and this variability could be as important as changes in the air-sea heat fluxes (W m-2). Changes in advective time lags are also linked to the width of the boundary current system, and thus to Bsurf. The boundary current was weak/broad in 1997 and 2003, coinciding with when you observe that largest temperature differences between GI and the BSO (Figure 8).
Suggestion: Revise your description to incorporate the above-described, more realistic scenario, including appropriate reference to Blindheim et al (2000).
Ensure that your Conclusions clearly state that you have identified: 1) reduced cooling north of Gimsøy and 2) established a link between this phenomenon and reduced sea-to-air heat fluxes. Additionally, 3) with a more realistic inclusion of the outer baroclinic currents and your (trimmed) advection-lag analysis, you could also explain the 2000-2005 and post-2017 periods. You should mention eddy fluxes, but in a more tentative manner.
Details
Lines 9-11: This wording may give the impression of temporal cooling. Please use better wording, similar to the description at the beginning of the Introduction (lines 21-24)
L27: “… while the more baroclinic NwAFC…”
L28: southern Norwegian location?
L34: bifurcates
L37: to à towards
L40: effect à affect
L81: Why do you only include data up to 2022
L214: Delete ‘figure’
L315-323: The year 2003 was characterized by very weak and broad boundary current system. This condition likely both reduced the advection speed, and increased the surface area of the AW (Bsurf). Inclusion of a temporally variable Bsurf could thus help you explain the anomalous signals around both 2003 (large Bsurf) and 2017 (smaller Bsurf).
Figures
Figure 1
Just out of curiosity, what do the negative (southward, blue) flows signify? There appear to be southward flows between the poleward AW current branches, with the highest values at the foot of the slop (flanking the slope current). Are these real, or artifacts from your combined altimetry-hydrography analysis?
Figure 7
Omit panel a)
Figure 8
Why not show the actual temperature differences between GI and the BSO, TGI - TBSO? These must, clearly, always positive (Figure 4). As presented, an incautious reader might interpret Figure 8 as indicating warming from GI to the BSO, after around 2011.
Citation: https://doi.org/10.5194/egusphere-2025-2854-RC1 -
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2854/egusphere-2025-2854-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
-
RC2: 'Comment on egusphere-2025-2854', Anonymous Referee #2, 03 Sep 2025
The paper is interesting, carrying an analysis of a massive series of repeated sections across the Norwegian Atlantic sloope Current, and the results seem to have been thoroughly investigated and supported.
I nonetheless have a series of comments and questions, that may be helpful to better grasp the main results.:
In AW current core for time series the top 50 m are removed (and thus also Ekman…), as well as constrain on maximum bathymetric depth (for SI and GI), in addition to constrain on S 35.00 (for practical salinity; for SA 35.16). What justifies not taking the top 50 meters? (or did I mis-understood?)
Then, the authors separate the core and non-core parts of this transport (based either on largest current or largest S, depending of the section to 50% of the transport…). I am not exactly sure of the practical interest to make this separation (except after to show that the results are someway independent on the part taken?). Upstream AW properties change, so what? Maybe it could be more straightforward to show the core+non-core, and then say (or in an appendix) that results are results hold in core (and non-core) that for the whole core+non-core (is that correct? I don’t think that this is shown or really discussed; indeed, in the abstract, it seems that there is more emphasis on what happens in the core, but after in the presentation, there is often a symmetry of results presented core and non-core)
Afterwards, even when describing the average core and non-core properties, one sees how each criteria (currents, bathymetry, salinity) strongly constrains the spatial extent of core and non-core; It seems to me that this might complicate a bit complicated the comparison of variations in different sections (in particular, between Svinoy and Gimsoy).
The main result (table 1 and figure 5) is that there is a decrease in the cooling between the sections off Norway and the northern sections, in addition to the multi-decadal variability found in all the time series (more for S than T, although in T I see it more in the non-core time series than in the core time series). The authors find reduced heat loss (sensible heat loss decreasing) and faster advection speed.
Question: what are uncertainties in the heat flux products and to which degree this can be taken at face value (maybe because it is on sensible component, and TA is better and more consistently reproduced in the reanalyses). I am not 100% sure that I fully understand the second result, which is not that clear to me in the time series from geostrophic velocity (and estimate of eddy transfer). Indeed, this interesting result is based not on the currents but on the lag correlations for salinity.
On Fig. 7, I did not fully understand how the vertical shear in the core is estimated and how it is relevant. In some places, cores defined by velocity, so it is important to specify in which layer this vertical shear is estimated: is the layer constant in time or varying? Is it also averaged across the core? Depending on how it is defined, the interpretation of the change is different (negative shear also means higher northward velocity at depth: is that correct?)
For interannual variability, figure 8 seems rather convincing (except for a little less than 10 years out of 31), with, if I understand correctly, advection maintained constant? I am no sure if trend removed? (I don’t think so, but that was not so clear to me)
Minor comments
- 34 Bifurcate
- 40: effect? Probably ‘result in’
- 100: if I understand the approach correctly, smoothing just on depths of isopycnals on individual section (not isopycnal T and S). What is done close to the sea surface? (or when an isopycnal outcrops?)? or near the boundaries (when isopycnal section crosses bottom…)
I understand what is done near bottom for extending T and S, and computing then Relative G V, but how is near-bottom transport then estimated? (I suspect taking exact bathymetry and multiplying by an area?)
And then, step of using at core value (0-error) the altimetry-derived velocity (l.113) to have absolute GV. This seems a reasonable approach. However, what are the errors (resolution; spatial smoothing of altimetric product? Is it similar to the one done on the isopycnal depths? )
- 204-209, I am wondering whether it would be better to present the trends difference Bjornoya-Svinoy an BSO-Svinoy, and not the other way around. Well, both can be argued… (saying than north warms more than south, thus less heat loss or other heat mechanism, seems to me a little bit easier to grasp, as water flows from south to north).
Heat fluxes, first l. 150, but after when the plots are shown, I would adopt the ocean convention as heat flux towards the ocean. Here seems that it is heat flux towards the atmosphere (as seen in figures). Thus, with the convention adopted here, an ocean heat loss is associated with a positive heat flux. I prefer the opposite, as is the vast majority of oceanographic papers that I looked at to verify.
- 258: ‘(Figure 7a,b)
Citation: https://doi.org/10.5194/egusphere-2025-2854-RC2 -
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2854/egusphere-2025-2854-AC1-supplement.pdf
-
EC1: 'Comment on egusphere-2025-2854', Meric Srokosz, 15 Sep 2025
If the reviewers' comments are addressed fully and convincingly in a revised manuscript this would become a publishable paper.
Citation: https://doi.org/10.5194/egusphere-2025-2854-EC1 -
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2854/egusphere-2025-2854-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Till Martin Baumann, 22 Oct 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 727 | 92 | 32 | 851 | 20 | 33 |
- HTML: 727
- PDF: 92
- XML: 32
- Total: 851
- BibTeX: 20
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Till M. Baumann
Øystein Skagseth
Randi B. Ingvaldsen
Kjell Arne Mork
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1921 KB) - Metadata XML
Review of Baumann et al.
This work gives valuable updates on transports and hydrographic variability, including long-term trends, at standard monitoring sections along the Norwegian slope. Focus is on reduced cooling at the northern part of this slope, which is most directly linked to declining air-sea heat fluxes. The manuscript is generally well written although too much detail is provided in places. Weaker/less robust aspects of this work (eddy fluxes, advection times) occupy too much space, and the apparent ignorance of the outer Norwegian Atlantic Front Current may be a severe limitation.
If the less robust parts are trimmed and the aforementioned severe limitation addressed (as suggested below), this could become a strong contribution to the literature.
Surface heat flux analysis
Throughout this manuscript, it is stated that the reduced cooling is due to reduced surface heat fluxes (especially sensible heat fluxes). This is introduced as early as line 50 (although disputed already in line 58). I suggest that this main result should be better emphasized, while that the weaker components (time lag and lateral eddy fluxes) should be downplayed.
I would structure the Results as follows:
Section 3.1 is well-executed, and should remain first. This section is, however, somewhat lengthy; some details could/should be moved to the Data & Methods part. The associated figure 3 and 4 are appropriate.
Section 3.2 makes an important contribution.
Section 3.3 discusses trends in temperature differences between the standard sections, relying only on references to Table 1, with no figure. The temporal variability in the temperature differences between the Gimsøy Section (GI) and the Barents Sea Opening (BSO) is not substantively presented in Section 3.4 and Figure 8.
Suggestion: Integrate the information in Section 3.4 into 3.3, so that Table 1 and Figure 8 provide mutual support.
Section 3.3.1 describes the most plausible driver; therefore give this sufficient emphasis.
Section 3.3.2 presents a rather heuristic attempt to discuss eddy fluxes. This is based on current meter moorings at the Svinøy Section, which is far upstream from the main signal (the reduced lateral cooling north of Gimsøy).
Suggestion: Move this section to your Discussion rather than presenting it as a primary result.
Section 3.3.3 is excessively long and detailed, without providing commensurate value to the paper.
Suggestion: Consider retaining this section in abbreviated form, showing only Figure 7b for the GI–BSO region, where your main signal is identified (e.g. Figure 8). The connections between the advective time lags and the discrepancy between the GI-BSO temperature difference and sea-to-air heat fluxes are not particularly strong. However, it is worth showing that the lags were short immediately after 2000 and after 2017, when the aforementioned discrepancy was large.
Inner and outer current branches
It is somewhat surprising that this work does not refer to Blindheim et al. (2000), whose abstract, amongst other findings, states: “A temperature rise in the narrowing Norwegian Atlantic Current is strongest in the north.”
This ‘narrowing’ refers to the outer branch (NwAFC), which is not addressed in the present work. This represents a potentially serious limitation. The spatial windows for estimating time series at the GI and SI sections represent a predominantly barotropic core/slope current. This is, however, not the case for the BSO, where the window is much broader, likely including influence from the outer branch, and represents much less of a (likely swift) slope current. Parts of the NwAFC passing through the Svinøy section, will turn eastwards with the topography of the Vøring Plateau, feeding the outer branch at GI; and likely also contributing to parts of the area selected for the BSO. The reported advection time lags of up to 12 months between the relatively closely positioned GI and BSO likely involve this outer current branch, rather than solely a fast nearly barotropic slope current, as portrayed here. This alternative (and in my opinion more realistic) perspective including the outer branch would require adjustments to your box model. Most importantly would the surface area of the Atlantic waters (Bsurf) become variable, and this variability could be as important as changes in the air-sea heat fluxes (W m-2). Changes in advective time lags are also linked to the width of the boundary current system, and thus to Bsurf. The boundary current was weak/broad in 1997 and 2003, coinciding with when you observe that largest temperature differences between GI and the BSO (Figure 8).
Suggestion: Revise your description to incorporate the above-described, more realistic scenario, including appropriate reference to Blindheim et al (2000).
Ensure that your Conclusions clearly state that you have identified: 1) reduced cooling north of Gimsøy and 2) established a link between this phenomenon and reduced sea-to-air heat fluxes. Additionally, 3) with a more realistic inclusion of the outer baroclinic currents and your (trimmed) advection-lag analysis, you could also explain the 2000-2005 and post-2017 periods. You should mention eddy fluxes, but in a more tentative manner.
Details
Lines 9-11: This wording may give the impression of temporal cooling. Please use better wording, similar to the description at the beginning of the Introduction (lines 21-24)
L27: “… while the more baroclinic NwAFC…”
L28: southern Norwegian location?
L34: bifurcates
L37: to à towards
L40: effect à affect
L81: Why do you only include data up to 2022
L214: Delete ‘figure’
L315-323: The year 2003 was characterized by very weak and broad boundary current system. This condition likely both reduced the advection speed, and increased the surface area of the AW (Bsurf). Inclusion of a temporally variable Bsurf could thus help you explain the anomalous signals around both 2003 (large Bsurf) and 2017 (smaller Bsurf).
Figures
Figure 1
Just out of curiosity, what do the negative (southward, blue) flows signify? There appear to be southward flows between the poleward AW current branches, with the highest values at the foot of the slop (flanking the slope current). Are these real, or artifacts from your combined altimetry-hydrography analysis?
Figure 7
Omit panel a)
Figure 8
Why not show the actual temperature differences between GI and the BSO, TGI - TBSO? These must, clearly, always positive (Figure 4). As presented, an incautious reader might interpret Figure 8 as indicating warming from GI to the BSO, after around 2011.