the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Feb 2026
Atlantic water intrusions onto the Scotian Shelf during the past 8.6 ka BP
Abstract. The Scotian Shelf lies at the confluence of warm Gulf Stream (GS) waters and the cold Labrador Current (LC), making it highly sensitive to large- and small-scale climate variability. Modern observations show rapid regional warming accelerated by episodic GS-derived intrusions, yet Holocene paleoceanographic reconstructions from this margin are sparse and often conflicting with respect to the frequency and extent of intrusion events. Here, we present high-resolution Mg/Ca-derived sea-surface temperature (SST) and planktonic δ¹⁸O records from St. Anns Basin on the north-eastern Scotian Shelf that provide new insights into the hydrographic surface-ocean variability of the past 8.5 ka calibrated Before Present (cal BP). While the SST record does not capture the 8.2 ka event, this event is evident in the δ¹⁸O and Ca/Sr records, indicating that its freshwater signal reached the Scotian Shelf. Reconstructed SSTs are generally cold from ~8.5 to ~6.2 cal ka BP, followed by a gradual increase in mean SSTs punctuated by multiple short-lived warm and saline events beginning around 6 cal ka BP, at 6.0–5.8, 5.5–5.4, 5.1–4.9, 3.2–3.1, 2.5–2.2 and 1.05–0.8 cal ka BP, which we interpret as intrusions of GS-sourced slope waters. We attribute these events to basin-scale reorganizations of the GS-LC system, consistent with the minimum/maximum modal state framework of Pickart et al. (1999). Minimum modal state circulation, characterized by a strong onshore LC and an intensified Deep Western Boundary Current (DWBC), which is dominated by Denmark Strait Overflow water, creates a sharp front which restricts intrusions of warm water onto the Scotian Shelf. Maximum modal state conditions feature a weakened LC and increased Labrador Sea Water (LSW) contribution to the DWBC, and reduce cross-slope temperature and salinity gradients that permit GS-derived waters to penetrate the shelf. Overall, our results indicate that warm-water intrusions occurred regularly throughout the past 6.5 ka BP with magnitudes of 6.7 °C and 1.5 psu comparable to those observed today.
-
Please read the editorial note first before accessing the preprint.
-
Preprint
(7477 KB)
-
Please read the editorial note first before accessing the preprint.
- Preprint
(7477 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-809', Anonymous Referee #1, 25 Mar 2026
- AC1: 'Author Comment on Revisions and Additions to the Manuscript', Henriette Marie Kolling, 28 Mar 2026
-
RC2: 'Comment on egusphere-2026-809', Anonymous Referee #2, 27 May 2026
Summary
This manuscript presents Mg/Ca-derived sea-surface temperature and planktonic δ¹⁸O records from a sediment core retrieved at 274 m water depth taken in the St. Ann's Basin on the northeastern Scotian Shelf. The authors reconstructed ~8.6 ka of hydrographic variability and argue that episodic warm and saline anomalies after ~6.5 ka represent Gulf Stream (GS)-sourced slope water intrusions, driven by modal-state shifts in the Gulf Stream–Labrador Current (GS-LC) system linked to AMOC, the Subpolar Gyre (SPG), and solar forcing. The topic is timely, and the site is well chosen, and it's a companion paper of Matzerath et al. (2026). However, several methodological, interpretive, and structural issues reduce confidence in the conclusions as currently framed.Major Concerns
1. Single-core, single-species design limits robustness
The entire reconstruction rests on one sediment core MSM101_44-3 and one foraminiferal species (Neogloboquadrina pachyderma). This species is known to exhibit a highly variable, ecologically sensitive depth habitat — the authors themselves acknowledge it calcifies at depths ranging from ~20 to over 100 m, depending on sea-ice cover, chlorophyll concentration, and stratification. With such a dynamic depth habitat, it is difficult to disentangle true SST signals from changes in the species' calcification depth over the 8.6 ka record. The authors argue that the reconstructed temperatures are consistent with modern subsurface conditions; however, this circularity (validating subsurface temperatures against subsurface observations while referring to them as "SSTs" in the title, abstract, and throughout the manuscript) is never fully resolved. The paper would be substantially strengthened by including a companion proxy less sensitive to depth migration — foraminiferal assemblage counts, dinocysts, or a second geochemical tracer on a different species — rather than relying on a single signal from a single organism.The authors' reliance on broader (including the Arctic) implications of using N. pachyderma rather than regional foraminiferal depth-habitat is either (1) willful ignorance or (2) disregard for other relevant data. The following three papers (including others) deal with foraminiferal assemblages in the southern Labrador Sea and Grand Banks!
Barrenechea Angeles, I., Lejzerowicz, F., Cordier, T., Scheplitz, J., Kucera, M., Ariztegui, D., Pawlowski, J., Morard, R., 2020. Planktonic foraminifera eDNA signature deposited on the seafloor remains preserved after burial in marine sediments. Scientific Reports 10, 20351.
Sahoo, N., Syed, M., Syed, S., Matul, A., Mohan, R., Tikhonova, A., Kozina, N., 2022. Planktic foraminiferal assemblages in surface sediments from the subpolar North Atlantic Ocean. Front. Mar. Sci. 8 https://doi.org/10.3389/fmars.2021.781675.
Stangeew, E. (2001): Distribution and isotopic composition of living planktonic foraminifera N. pachyderma (sinistral) and T. quinqueloba in the high latitude North Atlantic. PhD Thesis, Mathematisch-Naturwissenschaftliche Fakultät der Christian-Albrechts-Universität zu Kiel, 90 pp.2. The SSS reconstruction is insufficiently validated
Sea surface salinity is reconstructed from δ¹⁸O of seawater (δ¹⁸Osw) using a global mean ocean δ¹⁸O–salinity relationship (LeGrande & Schmidt, 2006). The Scotian Shelf sits at the confluence of isotopically distinct water masses — the cold, isotopically depleted Labrador Current and the warm, more enriched Gulf Stream — making a global mean equation a poor fit. Regional δ¹⁸O–salinity relationships for this margin differ substantially from the global mean, and the use of a single global equation to quantify salinity anomalies of only ~1.5 psu (the magnitude invoked to support intrusion events) is not adequately justified. Kolling et al. apply a global equation to a highly variable regional environment and then use the resulting SSS anomalies as primary evidence of intrusions. This is a notable weakness that should either be addressed through regional calibration or, at a minimum, treated with far more explicit uncertainty. Did the author consider the following relevant reference? Would the Figs. 3, 5, and 6 of Khatiwala et al. (1999) support Kolling et al.’s findings?Khatiwala, S. P., Fairbanks, R. G., and Houghton, R. W. 1999. Freshwater sources to the coastal ocean off northeastern North America: evidence from H218O/H216O. JGR 104 (C8), 18241-18256.
The author used the generic calibration equation of Elderfield and Ganssen (2000) to convert their Mg/Ca data despite numerous calibration equations specifically developed for N. pachyderma. For example, Kozdon et al. (2009), Nürnberg (1995), Wu and Hillaire-Marcel (1996, GCA); Livsey et al. (2020).
Kozdon, R., A. Eisenhauer, M. Weinelt, M. Y. Meland, and D. Nürnberg (2009), Reassessing Mg/Ca temperature calibrations of Neogloboquadrina pachyderma (sinistral) using paired δ44/40Ca and Mg/Ca measurements, Geochem. Geophys. Geosyst., 10, Q03005, doi:10.1029/2008GC002169.
3. Chronological resolution and uncertainty are underappreciated
The age model is based on excellent 10 previously published AMS ¹⁴C dates on mixed benthic foraminifera, yielding 95.4% credible interval widths of 350–530 years. The warm intrusion events the authors identify — which are core interpretive units lasting only 100–300 years — are therefore within or below the age model uncertainty. It is possible that two adjacent data points, identified as a single coherent "event," straddle an age-model uncertainty of several centuries. The manuscript does not propagate chronological uncertainties into the event timing or duration claims, and comparisons with other Holocene records (e.g., Bond events, AMOC minima reconstructions) are made with a precision the age model cannot support. A more rigorous treatment of age uncertainty, including Monte Carlo propagation or, at a minimum, a clear discussion of how age uncertainty affects event identification, is needed. Moreover, the calibration used by the authors is erroneous, as Heaton et al. (2023) revised their original “recipe” for the polar regions (see the reference below).Kolling et al. did not consider the impact of sea ice on modifying R. Their R data is not convincing. Sea ice and wind are not the only modifications of R, and different water masses, for example, Arctic-derived vs tropics will have a different R. Also, just because there is no sea ice influence at the site today doesn’t mean it could not have happened in the past. See Wanamaker et al. (2012).
Heaton, T. et al., 2023. Marine Radiocarbon Calibration in Polar Regions: A Simple Approximate Approach using Marine20. Radiocarbon 65 (4), 848 - 875.
Lewis, C.F.M. et al. 2012. Lake Agassiz outburst age and routing by Labrador current and the 8.2 cal ka cold event. Quat. Int. 260, 83–97.4. The event-detection methodology is ad hoc
Warm events are defined as data points exceeding one or two standard deviations above the mean SST and SSS. This is a common but statistically arbitrary approach — the threshold is not grounded in any physical or oceanographic criterion. Two single-point peaks are excluded without a fully providing statistical justification (the authors simply note they consist of "only one data point"). Moreover, the mean values used for thresholding (2.4°C and 32 psu) integrate both the cold early period and the warmer mid-to-late Holocene, which means the statistical threshold shifts depending on which part of the record is being evaluated. A more robust approach — such as a running mean baseline, a formal change-point detection algorithm, or comparison against a null hypothesis of red noise — would place the event identification on firmer ground.If an outlier analysis of the data plotted in Fig. B1 is carried out, six data points (621, 527, 503, 495, 340, and 335 cm) in Fig. B1 and subsequent curves C-F in Fig. 3 would be beyond the 95% confidence interval or would fall beyond the trend-line. Thus, the decadal-to-centennial-scale variability discussed in the manuscript would disappear.
5. The modal-state framework is applied somewhat loosely
The Pickart et al. (1999) modal-state framework was developed specifically for the Newfoundland Shelf and hinges on the relative contributions of Denmark Strait Overflow Water (DSOW) versus Labrador Sea Water (LSW) to the WBC. Kolling et al. invoke this framework for the Scotian Shelf, which lies further southwest and has a different shelf geometry, bathymetry, and water-mass exposure. The adaptation is plausible but is not rigorously tested. No proxy record from this core actually constrains WBC composition — the authors infer minimum versus maximum modal states entirely from the surface SST-SSS signal, which creates a degree of circular reasoning: the warm events are taken as evidence of maximum modal states, and the maximum modal state is then offered as the explanation for the warm events. Just a reminder that the sediment core used by Kolling et al. was retrieved at 274 m water depth bathed by the inner Labrador Current. Thus, the dynamics of the inner Labrador Current must be clearly linked to the LSW and WBC before the authors’ hypothesis can be accepted, as outlined in section 4.5. I am aware that the independent constraints on deep- or intermediate-water variability at this site (e.g., benthic foraminiferal δ¹³C or Cd/Ca) could not be achieved due to the unavailability of various proxy-carriers; however, rigorous testing is needed before accepting this mechanism rather than merely assuming it.6. Incomplete comparison with contradicting records
Kolling et el. briefly acknowledge that their Mg/Ca SST record contradicts alkenone-based reconstructions from the nearby Emerald Basin, which show a long-term cooling trend with no centennial variability. While the authors offer plausible arguments (seasonal bias, depth habitat, bloom seasonality), they do not systematically engage with why the same intrusion events identified at St. Ann's Basin leave no signal in the foraminiferal assemblages of Emerald Basin. The proposed explanation — differences in basin geometry and proxy sensitivity — is, to be precise, speculative and underdeveloped, and thus unacceptable. Given that Emerald Basin is central to the regional Scotian Shelf literature and provides the most directly comparable record, a more thorough and spatially explicit discussion of why proxy signals diverge across basins only ~100 km apart is essential.Minor Concerns
The title refers to "8.6 ka BP" but the substantive interpretive findings largely concern the last 6.5 ka. The early part of the record (8.6–6.5 ka) is characterized as stable and essentially featureless. Readers may reasonably expect more emphasis on the older interval.The Mn/Ca correction applied to all Mg/Ca values warrants more transparency. Kolling et al. conclude that corrected and uncorrected values do not differ; the fact that 62 of their samples exceed the Boyle (1983) threshold for diagenetic coatings is substantial and should be discussed more fully rather than given a brief note.
XRD determined Calcite (wt%); it would have been useful to obtain dolomite to identify the Lake Agassiz or 8.2 ka event robustly.
The spectral analysis section makes claims about solar forcing (the de Vries cycle and the 1500-year Bond cycle) based on a record spanning only ~8.6 ka and unevenly sampled. The spectral significance thresholds (95% confidence intervals) are reported, but the bandwidth and degrees of freedom in the Blackman-Tukey analysis are not, making it difficult to assess the robustness of the spectral peaks independently.
The SSS-to-δ¹⁸O conversion implicitly assumes no change in regional ice volume correction beyond the Siddall et al. (2003) global sea-level correction. For a near-coastal site influenced by variable glacio-isostatic adjustment and proximity to the retreating Laurentide Ice Sheet, this assumption should be explicitly acknowledged as a potential source of error. Could the authors not find reasonable references for the Holocene sea-level changes surrounding the North Atlantic than the far-field of the Red Sea? How about Walker et al. (2021)?
A lot is riding on the Wharton (2022) thesis in this manuscript; however, access to the thesis is restricted to UCL open-access staff until 1 March 2027. Therefore, we have to accept Kolling et al. statements at face value rather than verify them.
Walker, J.S., et al., 2021, Common Era sea-level budgets along the U.S. Atlantic coast: Nature Communications, v. 12, p. 1841, https://doi.org/10.1038/s41467-021-22079-2.
We thought the German research institutions had made progress in dissuading chain authorship; however, it appears to be a spurious attempt, as shown by the current authorship.
Overall Assessment
The paper addresses a genuinely important problem — the paleoceanographic history of a rapidly warming and poorly understood margin — and provides what is, to date, one of the few high-resolution foraminiferal records from the Scotian Shelf. The (poor) identification of episodic warm intrusions and their tentative linkage to North Atlantic circulation modes is a potentially significant contribution. However, in its current form, the study overreaches in several places: (1) the event-detection approach is underdeveloped, (2) the SSS reconstruction rests on an inappropriate calibration, (3) the claimed chronological precision in event timing exceeds what the age model supports, and (4) the mechanism invoked (modal-state transitions) is inferred rather than independently tested. Substantial revision is needed before the interpretations can be accepted with confidence. The paper is recommended for major revision or rejection.Citation: https://doi.org/10.5194/egusphere-2026-809-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,068 | 569 | 125 | 1,762 | 84 | 116 |
- HTML: 1,068
- PDF: 569
- XML: 125
- Total: 1,762
- BibTeX: 84
- EndNote: 116
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Henriette M. Kolling
Markus Kienast
Peter Matzerath
Julia Gottschalk
Stephanie Kienast
Daniel A. Frick
Felix Gross
Jack Wharton
David Thornalley
Ralph R. Schneider
Please read the editorial note first before accessing the preprint.
- Preprint
(7477 KB) - Metadata XML
The paper presents a new and fairly high resolution reconstruction (of Holocene SST, salinity and elements including calcite) from the Scotian Shelf and identifies several abrupt and short-lived warm episodes within the Holocene, interpreted as being incursions of gulf stream water onto the shelf. The introduction, study area, methods and results are concise and well written, and the comments below related to these sections are minor. The data-set is really interesting!
I recommend below that the discussion is in places restructured/re-written to provide a more succinct and clear explanation, because some ideas and interpretations are spread over different paragraphs and sections. I also have some questions about some of the interpretations (particularly on the argument that the NAO would not have been an influence), although I agree with most of the authors interpretations (e.g. the gulf stream being the source of the shelf variability, the early Holocene influence of melt water).
Abstract:
Sentence at line 17-19 - mention within this sentence that salinity is reconstructed
Line 20 - 'and Ca/Sr records' - first mention of these - say that this proxy has been developed in preceding sentence
line 28 - 'and reduce cross-slope' - reword to 'which reduces cross-slope...' perhaps
line 29 - 'warm-water', given the rest of the sentence you should say here warm, saline water...'
Introduction:
Line 38-39 - how much of a warming trend has there been? A detail highlighting the scale would be good here. In addition detail about the temperatures achieved during the warm intrusions compared to the average trend, as well as the duration of the intrusions would be good. For example saying ‘GS intrusions since 2011 have increased SST's from X to X for periods of X weeks’.
Line 48 - think this should be 'from' rather than 'on'
Line 65 - 'additional freshwater reconstruction', the wording here is unusual. Do you mean a salinity reconstruction?
Study Area:
Line 92 - do you mean 'during the last 50 years'? rather than 'for more than 50 years'?
Materials and Methods:
Methods - the descriptions of what you did is well written and thorough. Through the methods I would like to see a bit more explanation of why certain methods are being done. For example sentences at the start of the subsections saying e.g.' XRD analysis was applied to determine X, Y, and Z which are proxies for X (+ references)'. It otherwise assumes that the reader is familiar with what different proxies are used for and doesn’t give any context of previous studies that have led to these methods being used.
Line 103-104 - What was the sieving for? Its not clear here currently, link it to the proxies.
line 155 – typo - specimens
Results:
line 207 - typo ' complete'
Line 212 - it should read 'range between 0.7 and 2.8' I think
line 230 - the wording here is unclear. Not clear if the decreasing trend is through the whole record, and if this is the case why mention 6ka? I think you mean it gradually decreases until 6 ka and then a steeper decrease occurs towards the core top?
Discussion:
Line 236 - the warming trend is not clear at all from looking at the figure. Is it significant and worth mentioning? I would almost say it was warmer between 6 and 2 ka BP but cooler before and after, but with no longterm trend.
Line 238 - do you mean the long term trends differ or the presence/timing of warm anomalies differ?
line 241 - warming trend or warm period? please clarify
Lines 244-247 - could another explanation be a stronger GS associated with a stronger SPG circulation and greater cold water transported southwards along the LC?
Line 249 - the first 2 paragraphs are quite confusing, jumping between proxy explanations, physical explanations, warming/cooling trends in different areas... consider if these could be restructured or clarifying sentences added to guide the discussion
line 255 - rather than 'here' say scotian shelf, otherwise could mean this paper
line 257 - 'in contrast' - do you mean in contrast to the suggestion of Osman about frontal migration? Perhaps say that 'In contrast, other studies of alkenone records highlight'
Line 261 - I am not an expert in the proxies, but wonder if changes in seasonal SST's were influencing the alkenone records how do you know that seasonal SST shifts are not having an influence on your record?
Line 266 - this paragraph doesn’t fit well here. I think this point should have been made before the above discussion of the SST reconstruction at the start of this section
Line 272 - in this paragraph your record shows stable but cool SSTs at 8.5-6.5, and you go on to say this is the same as records with cooling trends? Please be clear in your description whether you mean these other records had cooler temperatures during this period or cooling trends, as these are not the same thing
Line 279 - here you say your record shows a cooling, not stable temperatures which is a contradiction
Line 288-289 - while the final sentence here leads into the next section I don’t think it is needed here. I would include it (if necessary) at the start of the next section
Line 318 - could the move to deeper waters have also isolated the foraminifera from atmospheric changes in temperature, resulting in the lack of a signal?
Line 324 - this should be part of the above paragraph I think.
Line 329 - This paragraph seems to be a summary of the above discussion. If that is the case then it should be highlighted (e.g. In summary, between...' Otherwise I would remove this paragraph as a bit repetitive of the above section
Line 350 - 'this interpretation contrasts with...' I am not sure there is a contradiction here, as you would expect conditions on the shelf to be influenced by different conditions to the NW Atlantic, which covers a much larger area of open ocean. Instead of implying there is a disagreement or debate here you could highlight the shelf and open ocean conditions differed, and then say the warm signal is fairly muted in your record.
Line 360 - do you mean re-expansion or just expansion?
Line 361 - as strong stratification has not been directly measured by your study I would suggest making this statement more cautious. '...strong stratification that likely characterized...'
Line 363 - the separate section on the GS intrusions below was unexpected as I considered that the short term variability in the record was discussed in this section. Perhaps add some signposts for the reader, such as highlighting at the start of section 4.3 that this section is looking at the broad oceanography over time and milllennial changes and the following section is more local or centennial oceanography. Such signposts would provide some justification for discussing the variability in section 4.3 and 4.4. An alternative would be to combine sections 4.3 and 4.4.
Line 368-369 - the results of Levac are not inconsistent with yours, they support yours as summer SST's were low?
Line 423-428 - looking at figure 5 the Shuman drought periods described here do not align with the warm or cold periods in your SST reconstruction. I am also not sure that the mechanisms here support or bring much to your description of the mechanisms, but perhaps it is just that more explanation is needed. Are you suggesting that atmospheric changes were driving the gulf stream incursions? Or that the GS incursions contributed to the drought/wet phases in America?
Line 435 - 'reflects wintertime atmospheric variability...operated on interannual to decadal timescales', I am not sure that this is the case and that the influence of the NAO should be disregarded for these reasons. The observations we have show just annual/decadal changes because they are short, palaeoclimate reconstructions of the NAO (e.g. Trouet et al. 2009; Olsen et al., 2012) show centennial variations. There is also literature on the NAO influence on GS strength and position. While the NAO is a winter phenomenon the influence of wind forcing on ocean circulation continues through the year. Similarly winter atmospheric temperature anomalies could influence sea ice thickness, persistence and thus spring/summer conditions.
Line 457 - I disagree as above that the impacts of the NAO are only felt in winter and have no impact on spring and summer conditions. The SPG and AMOC strength you have highlighted as important controls, and these are driven by wind stress and changes in winter deepwater formation. These therefore provide several ways that NAO variability can have an influence on the ocean year around - winter temperatures, wind and rainfall over the deepwater formation regions (influenced at least in part by the NAO state) would control NADW formation rates and thus the AMOC strength. Winter wind stress over the Atlantic would control the surface currents and SPG. These responses would not just be during winter, so would influence the GS intrusions onto the scotian shelf.
Line 468 – have ‘Bond Events’ been solely linked to solar variability? Is it not also linked to internal ocean variability? If there is debate about this I would not highlight it as a proxy for solar variability in Figure 5.
Line 473 - I would suggest merging this section on spectral analysis with the above discussion on the drivers of the GS intrusions. While the discussion here stems from the cycles identified, it is providing a more in depth discussion of the processing that were started in the above section. Combining it with the above section, and just mentioning the spectral results there, would strengthen the explanation I feel.
Line 486 - I am not sure what the resolution is of the record, but if it is over 100 years I would caution against reading too much into the 200-300 year cycles here.
Conclusions:
Line 512 - mention the other proxies, calcite and sea surface salinity as well perhaps?
Figure 4 - it is not clear from the plot or the caption which is which in plot D (the Wharton and Sachs records)
Figure 5 - the event around 4.7 ka BP does not have a shaded bar, its similar magnitude to other events.