the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Aug 2024
Postglacial environmental changes in the northwestern Barents Sea caused by meltwater outbursts
Abstract. The last deglaciation was marked by abrupt shifts between cold and warm states reflecting an integrated response to the gradually increasing summer insolation at northern latitudes, changing ocean circulation, and the retreat of the Northern Hemisphere ice sheets. In this study, we present new multiproxy reconstructions of water mass properties and sea surface characteristics from a sediment core from the northwestern Barents Sea (Kveithola) representing the last 14,700 years. Our reconstruction documents four sediment-laden meltwater pulses between 14,700 and 8,200 cal years BP based on biomarkers, stable isotopes, and sedimentological parameters. Deglacial processes primarily cause these meltwater pulses and are possibly supplemented with paleo-lake outbursts, paleo-tsunami currents, or a combination of at least one of these, are characterized by sudden drops in sea surface temperatures, increased sea ice formation, increased terrigenous supply, and a limited influence of Atlantic Water in the northwestern Barents Sea. The influence of the Storegga tsunami, which occurred around the 8,200 cal years BP cooling event likely reached and redistributed the sediment in Kveithola. Strong coarsening of the northwestern Barents shelf was observed after 3,500 years, which might be related to a stronger Atlantic Water inflow from the west across the bank leading to winnowing.
- Preprint
(2609 KB) - Metadata XML
-
Supplement
(326 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on cp-2024-52', Anonymous Referee #1, 21 Aug 2024
The paper by Devendra et al. “Postglacial environmental changes in the northwestern Barents Sea caused by meltwater outbursts ” attempts to reconstruct the paleoenvironment and the depositional regime from a single core from the trough Kveithola which is located at the western Barents Sea shelf and dates back to about 14.5 ka. To do so they employ quite a number of different proxy data. These include studies in micropalaeontology (benthic and planktonic foraminiferal assemblages and their O-isotopic composition), sedimentology (grain size fractions), geochemistry (XRF), and a suite of biomarker tools. They use all of these data to focus on 3 events, which they identify to be meltwater-triggered, and then, eventually, compare/relate the youngest of those to overregional changes far away.
While the paper is well written, its main focus, namely reconstructing meltwater outbursts during the first ~5ky of the Holocene from these proxies, is fundamentally flawed and therefore not acceptable for publication. The main question, apart from a solid age framework (see below), is to what extent the different proxies add up and can provide a consistently coherent story that also delivers some concrete information on the actual causes of their meltwater events? Instead the authors remain rather vague on this issue, naming all sorts of processes, from far-distance palaeolake drainage of the Baltic basin to tsunami-triggered currents by the Storegga slide to sediment-laden meltwater plumes....
Age model and associated problems: The 142cm long core has 10 radiocarbon dates, the lowermost 2 being of similar age. Unfortunately for the interpretation there are also 2 reversals, the lower one covers a huge age range (~4ky), and with 50cm amounts to more than 1/3 of the entire core length. And still, all proxy data from this section are being interpreted, including the YD interval and their MP1.
The second reversal at depth 53cm is interesting, because of its very old radiocarbon age which deviates from the dates below and above by about 15ky. It indicates that the bulk of forams measured were mostly very old, perhaps 30-40ka, and an admixture of just a few much younger ones caused the final age. The crucial thing here is that depth 53 cm is in dead center in their MP3 (around their 8.2ka event), meaning whatever the cause, there was plenty of reworking and delivery of older sediments to the site, including all the fine TOC-rich sediments that would contain their biomarker proxies. The section of MP3 between the 2 acceptable dates is 20 cm long, telling us, together with the first reversal, that half of the 142cm long core (70cm) essentially remained undated. Moreover, the XRF-data (Fig. 4) across MP3 seems to imply that a larger sediment section was affected than indicated by the blueish bar.
Speaking of the biomarkers, in the polar ocean elevated tetra-unsaturated alkenones (Uk37:4) are traditionally indicative for freshened waters and admixture of sediments which both could be derived from ice rafting. However, what is very critical using these samples with high 37:4 to be cautious when calculating SSTs from the 37:2 and 37:3. Indeed, the SSTs do not provide any consistency for huge temperature drops as claimed by the authors.
Foraminifers: The authors claim that the entire core is dominated by planktonics...but my pdf only shows their occurrence after 8ka – incomplete because something wrong with Fig.4? The Barents Sea is known to be prone to massive calcite dissolution and I wonder of the authors noted some drastic changes throughout the core.
The oxygen stable isotopes measured on the foraminifers appear very inconclusive with no proper trends over the entire core; the planktonic are only based on 15 samples and measure after MP3. The benthics vary over a 1 permil range and show no relation to any of the 3 MPs in terms of O-depletions. This is somewhat surprising considering that the authors also talk about sediment-laden plumes associated with these meltwater events.
Sedimentology: I am puzzled about the use of grain sizes. In Fig. 4 the Coarse Fraction (>63µm), which to my mind is representative of the sand size fraction, does not exceed 5-7%. In Suppl-Fig, there is a completely different “sand” curve shown....
Concerning the figures in general there were no plots that would give the reader a better insight and understanding of the main proxies vs. depth (not even in the suppl.). Instead the authors provide all figures vs. the final age model with a rather obscure floating depth scale below.
Besides of all the above, I do appreciate the effort of the authors by compiling the various proxy records. And yet, a large amount of data/publications is available from well-dated sediment cores downstream from their core position. It is somewhat unfortunate to realize that the authors do recognize some of these cores on the map (e.g. GIK23258 and farther north) but fail to make any further direct and visual comparisons with these excellent Holocene proxy records.
Citation: https://doi.org/10.5194/cp-2024-52-RC1 -
AC1: 'Reply on RC1', Dhanushka Devendra, 09 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Dhanushka Devendra, 09 Oct 2024
-
RC2: 'Comment on cp-2024-52', Anonymous Referee #2, 09 Oct 2024
In this manuscript, entitled as “Postglacial environmental changes in the northwestern Barents Sea caused by meltwater outbursts”, you tried to detect events of meltwater outburst and/or paleo-tsunami in the northwestern Barents Sea during the last deglaciation period. Your results by using multi-proxies are enough to explain phenomena, but the descriptions for the discussions and conclusion are still unclear.
Major correction
- Introduction
You mentioned several proxies to clarify your evidences. However, you didn’t deeply explain about proxies and examples of their usages. If you can, please add these descriptions to Introduction (or Discussions). Otherwise, readers cannot follow your discussion anymore.
- Organic geochemical analyses.
You used the response factor provided by Belt et al. (2013) for biomarker analysis. However, response factor is different between the machine condition/setting of GC-MS. You used exactly same machine and method to analyze it. If not, it is better to analyze IP25 using GC for making your own response factor to calculate concentrations from GC-MS data.
There is no introduction for biomarkers such as alkenone (especially C34:4), IP25 and steroids. It seems that you choose more critical indicators among several biomarker proxies. However, the reasons to choose those indicators are still unclear for reader. Especially, you choose C34:4 and PBIP25. Please carefully explain why you choose them for paleoenvironment description.
Citation: https://doi.org/10.5194/cp-2024-52-RC2 -
AC3: 'Reply on RC2', Dhanushka Devendra, 28 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC3-supplement.pdf
-
AC4: 'Reply on RC2', Dhanushka Devendra, 28 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC4-supplement.pdf
-
EC1: 'Comment on cp-2024-52', Stephen Obrochta, 09 Oct 2024
Based on my own reading of the manuscript, I tend to agree with the assessment of Reviewer 1 and find the author's response to be lacking. While I commend the authors for employing a wide variety of analyses, I am also struggling to see how these data are supporting the interpretation of meltwater pulses. Data resolution in general is problematically low for the majority of the datasets. The methods state the core was sampled at a 1-cm interval, which should result in nearly 150 samples. However the figures appear to show <50 data points for the biomarker data. The methods do not mention the resolution of the analyses. In my opinion, a much higher resolution dataset would be more convincing.
I strongly suggest that, at the very least, the manuscript be better organized and with more detailed explanations. I am finding parts that should be in the methods, such as the description of the age-depth modeling routine that was used, in the results. Conversely, the result of the age-depth modeling is contained in the methods. The parameterization of Bacon should also be included in the methods. Figure 3 in the results includes interpreted meltwater events, but these have yet to be introduced in the text. The ordering of information presentation is important. As is, the manuscript is difficult to follow. The term "AW" is never defined in the main text, only in a figure caption after its first usage.
The methods do not fully explain the methods that were used. For example, section 3.1 (line 110) says the samples were "wet sieved through 100 and 500 µm meshes". This is of course non-standard, and I assume this was not done. Figure 4 shows the >63 µm fraction, so either the statement on line 110 is incorrect, or the corresponding y-axis text label in Figure 4 is incorrect. In Section 3.3 (line 145) dry bulk density is mentioned but there is no description of how it was measured (pycnometer?) In Section 3.4 (line 160), the manuscript states that IRD counts were performed on the >500 µm fraction. IRD counts are typically performed on the >150 µm fraction and using a non-standard size won't allow comparison to other IRD records. Furthermore, the vast about of IRD in icebergs is in the fine fraction, so the statement that melting icebergs "can be ignored" is not really supported (line 311). I also believe that the raw data output from the Olympus Vanta M need to be converted to oxides before interpretation, but the methods (Section 3.6) make no mention of this. Was this done? If so, how was it done? The raw output element trends can be very similar without conversion and will vastly change after oxide calculation. Therefore, the similarity between Fe and Ti may be coincidental. The methods must completely and accurately describe all analyses performed in the paper, and what is described gives me serious reservations about the resulting data.
All age modeling routines assume that the age determinations reflect the age of sediment deposition. The age reversals, as pointed out by Reviewer 1, do indeed provide evidence of sediment reworking, which in turn does subject the resulting modeled median age to additional uncertainty. I further think that any future version of the manuscript should contain the uninterpreted data plotted versus core depth. I don't see how the figure included in the response to Reviewer 1 supports the statement that "The trends in the multi-proxy records are consistent between the cores." The only two proxies that are the same between the two cores are brassicasterol and SST (why are the other records even shown?), and the trends within the undated sections do not appear similar to me. Brassicasterol absolute values are two orders of magnitude higher in the other core and exhibit a decreasing trend, while there is essentially no trend in the data of the present manuscript. In the SST records, I believe the authors may be referring to the two drops at ~11.5 and ~12.5 ka, but due to the coarseness of the data, it cannot be ruled out that these drops are simply outliers. Hence, my initial comment on increased data resolution.
Citation: https://doi.org/10.5194/cp-2024-52-EC1 -
AC2: 'Reply on EC1', Dhanushka Devendra, 28 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC2-supplement.pdf
-
AC2: 'Reply on EC1', Dhanushka Devendra, 28 Oct 2024
Status: closed
-
RC1: 'Comment on cp-2024-52', Anonymous Referee #1, 21 Aug 2024
The paper by Devendra et al. “Postglacial environmental changes in the northwestern Barents Sea caused by meltwater outbursts ” attempts to reconstruct the paleoenvironment and the depositional regime from a single core from the trough Kveithola which is located at the western Barents Sea shelf and dates back to about 14.5 ka. To do so they employ quite a number of different proxy data. These include studies in micropalaeontology (benthic and planktonic foraminiferal assemblages and their O-isotopic composition), sedimentology (grain size fractions), geochemistry (XRF), and a suite of biomarker tools. They use all of these data to focus on 3 events, which they identify to be meltwater-triggered, and then, eventually, compare/relate the youngest of those to overregional changes far away.
While the paper is well written, its main focus, namely reconstructing meltwater outbursts during the first ~5ky of the Holocene from these proxies, is fundamentally flawed and therefore not acceptable for publication. The main question, apart from a solid age framework (see below), is to what extent the different proxies add up and can provide a consistently coherent story that also delivers some concrete information on the actual causes of their meltwater events? Instead the authors remain rather vague on this issue, naming all sorts of processes, from far-distance palaeolake drainage of the Baltic basin to tsunami-triggered currents by the Storegga slide to sediment-laden meltwater plumes....
Age model and associated problems: The 142cm long core has 10 radiocarbon dates, the lowermost 2 being of similar age. Unfortunately for the interpretation there are also 2 reversals, the lower one covers a huge age range (~4ky), and with 50cm amounts to more than 1/3 of the entire core length. And still, all proxy data from this section are being interpreted, including the YD interval and their MP1.
The second reversal at depth 53cm is interesting, because of its very old radiocarbon age which deviates from the dates below and above by about 15ky. It indicates that the bulk of forams measured were mostly very old, perhaps 30-40ka, and an admixture of just a few much younger ones caused the final age. The crucial thing here is that depth 53 cm is in dead center in their MP3 (around their 8.2ka event), meaning whatever the cause, there was plenty of reworking and delivery of older sediments to the site, including all the fine TOC-rich sediments that would contain their biomarker proxies. The section of MP3 between the 2 acceptable dates is 20 cm long, telling us, together with the first reversal, that half of the 142cm long core (70cm) essentially remained undated. Moreover, the XRF-data (Fig. 4) across MP3 seems to imply that a larger sediment section was affected than indicated by the blueish bar.
Speaking of the biomarkers, in the polar ocean elevated tetra-unsaturated alkenones (Uk37:4) are traditionally indicative for freshened waters and admixture of sediments which both could be derived from ice rafting. However, what is very critical using these samples with high 37:4 to be cautious when calculating SSTs from the 37:2 and 37:3. Indeed, the SSTs do not provide any consistency for huge temperature drops as claimed by the authors.
Foraminifers: The authors claim that the entire core is dominated by planktonics...but my pdf only shows their occurrence after 8ka – incomplete because something wrong with Fig.4? The Barents Sea is known to be prone to massive calcite dissolution and I wonder of the authors noted some drastic changes throughout the core.
The oxygen stable isotopes measured on the foraminifers appear very inconclusive with no proper trends over the entire core; the planktonic are only based on 15 samples and measure after MP3. The benthics vary over a 1 permil range and show no relation to any of the 3 MPs in terms of O-depletions. This is somewhat surprising considering that the authors also talk about sediment-laden plumes associated with these meltwater events.
Sedimentology: I am puzzled about the use of grain sizes. In Fig. 4 the Coarse Fraction (>63µm), which to my mind is representative of the sand size fraction, does not exceed 5-7%. In Suppl-Fig, there is a completely different “sand” curve shown....
Concerning the figures in general there were no plots that would give the reader a better insight and understanding of the main proxies vs. depth (not even in the suppl.). Instead the authors provide all figures vs. the final age model with a rather obscure floating depth scale below.
Besides of all the above, I do appreciate the effort of the authors by compiling the various proxy records. And yet, a large amount of data/publications is available from well-dated sediment cores downstream from their core position. It is somewhat unfortunate to realize that the authors do recognize some of these cores on the map (e.g. GIK23258 and farther north) but fail to make any further direct and visual comparisons with these excellent Holocene proxy records.
Citation: https://doi.org/10.5194/cp-2024-52-RC1 -
AC1: 'Reply on RC1', Dhanushka Devendra, 09 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Dhanushka Devendra, 09 Oct 2024
-
RC2: 'Comment on cp-2024-52', Anonymous Referee #2, 09 Oct 2024
In this manuscript, entitled as “Postglacial environmental changes in the northwestern Barents Sea caused by meltwater outbursts”, you tried to detect events of meltwater outburst and/or paleo-tsunami in the northwestern Barents Sea during the last deglaciation period. Your results by using multi-proxies are enough to explain phenomena, but the descriptions for the discussions and conclusion are still unclear.
Major correction
- Introduction
You mentioned several proxies to clarify your evidences. However, you didn’t deeply explain about proxies and examples of their usages. If you can, please add these descriptions to Introduction (or Discussions). Otherwise, readers cannot follow your discussion anymore.
- Organic geochemical analyses.
You used the response factor provided by Belt et al. (2013) for biomarker analysis. However, response factor is different between the machine condition/setting of GC-MS. You used exactly same machine and method to analyze it. If not, it is better to analyze IP25 using GC for making your own response factor to calculate concentrations from GC-MS data.
There is no introduction for biomarkers such as alkenone (especially C34:4), IP25 and steroids. It seems that you choose more critical indicators among several biomarker proxies. However, the reasons to choose those indicators are still unclear for reader. Especially, you choose C34:4 and PBIP25. Please carefully explain why you choose them for paleoenvironment description.
Citation: https://doi.org/10.5194/cp-2024-52-RC2 -
AC3: 'Reply on RC2', Dhanushka Devendra, 28 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC3-supplement.pdf
-
AC4: 'Reply on RC2', Dhanushka Devendra, 28 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC4-supplement.pdf
-
EC1: 'Comment on cp-2024-52', Stephen Obrochta, 09 Oct 2024
Based on my own reading of the manuscript, I tend to agree with the assessment of Reviewer 1 and find the author's response to be lacking. While I commend the authors for employing a wide variety of analyses, I am also struggling to see how these data are supporting the interpretation of meltwater pulses. Data resolution in general is problematically low for the majority of the datasets. The methods state the core was sampled at a 1-cm interval, which should result in nearly 150 samples. However the figures appear to show <50 data points for the biomarker data. The methods do not mention the resolution of the analyses. In my opinion, a much higher resolution dataset would be more convincing.
I strongly suggest that, at the very least, the manuscript be better organized and with more detailed explanations. I am finding parts that should be in the methods, such as the description of the age-depth modeling routine that was used, in the results. Conversely, the result of the age-depth modeling is contained in the methods. The parameterization of Bacon should also be included in the methods. Figure 3 in the results includes interpreted meltwater events, but these have yet to be introduced in the text. The ordering of information presentation is important. As is, the manuscript is difficult to follow. The term "AW" is never defined in the main text, only in a figure caption after its first usage.
The methods do not fully explain the methods that were used. For example, section 3.1 (line 110) says the samples were "wet sieved through 100 and 500 µm meshes". This is of course non-standard, and I assume this was not done. Figure 4 shows the >63 µm fraction, so either the statement on line 110 is incorrect, or the corresponding y-axis text label in Figure 4 is incorrect. In Section 3.3 (line 145) dry bulk density is mentioned but there is no description of how it was measured (pycnometer?) In Section 3.4 (line 160), the manuscript states that IRD counts were performed on the >500 µm fraction. IRD counts are typically performed on the >150 µm fraction and using a non-standard size won't allow comparison to other IRD records. Furthermore, the vast about of IRD in icebergs is in the fine fraction, so the statement that melting icebergs "can be ignored" is not really supported (line 311). I also believe that the raw data output from the Olympus Vanta M need to be converted to oxides before interpretation, but the methods (Section 3.6) make no mention of this. Was this done? If so, how was it done? The raw output element trends can be very similar without conversion and will vastly change after oxide calculation. Therefore, the similarity between Fe and Ti may be coincidental. The methods must completely and accurately describe all analyses performed in the paper, and what is described gives me serious reservations about the resulting data.
All age modeling routines assume that the age determinations reflect the age of sediment deposition. The age reversals, as pointed out by Reviewer 1, do indeed provide evidence of sediment reworking, which in turn does subject the resulting modeled median age to additional uncertainty. I further think that any future version of the manuscript should contain the uninterpreted data plotted versus core depth. I don't see how the figure included in the response to Reviewer 1 supports the statement that "The trends in the multi-proxy records are consistent between the cores." The only two proxies that are the same between the two cores are brassicasterol and SST (why are the other records even shown?), and the trends within the undated sections do not appear similar to me. Brassicasterol absolute values are two orders of magnitude higher in the other core and exhibit a decreasing trend, while there is essentially no trend in the data of the present manuscript. In the SST records, I believe the authors may be referring to the two drops at ~11.5 and ~12.5 ka, but due to the coarseness of the data, it cannot be ruled out that these drops are simply outliers. Hence, my initial comment on increased data resolution.
Citation: https://doi.org/10.5194/cp-2024-52-EC1 -
AC2: 'Reply on EC1', Dhanushka Devendra, 28 Oct 2024
The comment was uploaded in the form of a supplement: https://cp.copernicus.org/preprints/cp-2024-52/cp-2024-52-AC2-supplement.pdf
-
AC2: 'Reply on EC1', Dhanushka Devendra, 28 Oct 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 427 | 97 | 17 | 541 | 23 | 6 | 6 |
- HTML: 427
- PDF: 97
- XML: 17
- Total: 541
- Supplement: 23
- BibTeX: 6
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1