The authors would like to thank the anonymous reviewer for taking the time to review the manuscript and provide very constructive comments and suggestions.

We are happy with the positive evaluation, and agree to all comments and suggestions, and have changed the manuscript accordingly.

Our response to minor comments can be found in the file attachment.

Dear Marc,

Thank you for your very constructive review! We plan to reply in full to the reviewers at the end of the open discussion, but it seems that it could be beneficial to deal with the issue of the unpublished (in review) Svennevig et al. (2023b) manuscript at this stage.

The unpublished manuscript Svennevig et al. (2023b) can be made available to you. Then you can check the statements and features in the present Korsgaard et al. manuscript (under review here) which are dependent on the Svennevig et al. manuscript.

Would  you be interested in this?

We hope that the Svennevig et al. manuscript will be accepted and have a proper doi for the reference list before proceedings here are done.

On behalf of the authors,

Niels Korsgaard

We are pleased with the positive review. The authors would like to thank you for taking your time to review the manuscript and provide very constructive comments and suggestions, which helped improve the manuscript.

We agree that there are several examples of careless language and welcome that these are pointed out so that imprecise language and ambiguities can be removed or cleared up.

We agree to all major and minor comments and suggestions and will change the manuscript accordingly. One change that will also affect the title of the manuscript, is that “giant” will be replaced with “very large” when characterizing tsunami magnitude. See our comments to the minor reviewer comments for further detail.

Indeed, the manuscript in review here cannot stand on its own and relies on unpublished data from Svennevig et al. (2023b) which is now accepted for publication with no changes of relevance to the manuscript under review here in NHESS. It is virtually the same manuscript which we forwarded a copy of to you.

RC3: General comments (major RC3 comments)

RC3: I got very excited when I was asked to review this paper about tsunami deposits in lakes in Greenland. The photos of two of the cores show very nice examples of tsunami deposits in lakes, its facies are similar to the Storegga tsunami facies in coastal lakes along the Norwegian coast. However, after reading through the manuscript I ended up being disappointed. My main criticism is the field work. The work in the field has not been a thorough investigation – only one core from each lake basin!! The possible tsunami deposits found in two of the lakes are interesting and I think deserves a better study. I am also not convinced that there are two tsunami layers as the authors claim – it might just be that the two layers belong to one event, but this is difficult to solve without better mapping of the deposits in the field.

AC: Thank you for thorough review which has helped improve the manuscript substantially. Our response to your review is in blue font.

We agree that the study would have benefitted from more comprehensive fieldwork and would have made it more conclusive. Working with the available data, we will revise the manuscript to make these weaknesses and implications apparent to the reader, and in this way address the concerns of the reviewer.

Our response in point-by-point:

1. Field work/single core per lake.
   
   It is normal procedure in Greenland lake coring is to bring a single core (or site) per lake back to the lab for detailed analysis and fieldwork equipment and logistics are usually “light”. In addition to this, reconnaissance using a Russian corer is often, but not always, done in the field. Mapping of the complete lake stratigraphy (with transect) seems to be rarer than a reconnaissance. In a lone example, Wagner et al. (2006), a single core is used to interpret a layer as from the Storegga tsunami.
   
   
   
   That we did not do this means we are not able to test one of the criteria for identification of a tsunami layer. We do meet other criteria and we also eliminate other processes which could create a stratigraphy which could be mistaken for a tsunami layer. In addition, we point to a likely source of tsunami waves – the very large Holocene rock avalanche deposits found at the bottom of the Vaigat strait. We consider this strong but not perfect evidence of tsunami invasion of the Saqqaq lakes.
   
   
   
   
2. One or two tsunami layers.
   
   We agree with the reviewer, that if we had mapped the lake stratigraphy, it would most likely have been possible to be conclusive if there are one or two tsunami layers in the lakes. However, we maintain that the most likely explanation is that there are two layers, T1 and T2. We base this on the available data; the apparent limited erosion at the base of T2; the 14C ages below, above and inside the T2 tsunami layer are quite close; and although it is possible that a 42 cm thick monolithic block of horizontally laminated gyttja could be a rip-up clast, it seems more likely to represent the time between T1 and T2.
   
   
   
   We will address this concern by making the reader aware of the real possibility that there is only one single tsunami layer in the revised manuscript.
   
   
   
   
3. Radiocarbon dating / What is the age of the T1 tsunami layer?
   
   We agree with the reviewer’s concern, and he has been extremely helpful on this subject, specifically pointing out the issues with using mosses (aquatic bryophytes) and bulk sediments for dating, and how type of dated material and stratigraphical context is important. This has led to a revision of the age of T1, which makes the dates and facies consistent across the cores. It is interpreted as a unit representing one or more tsunamis occurring in the 5.8 to 7.2 cal. ka time bracket. As the reviewer think, the tsunami layers found by Long et al. (2015) in a lake in south Disko Bugt falls within this bracket and could be related.
   
   
   
   
4. Fieldwork/Impenetrable substrate/short cores.
   
   We expected to get much longer cores. However, we were unable to penetrate what we during fieldwork subjectively judged to be bedrock or a boulder layer. At the same time, we had an issue with coring tubes that broke at the end and sometimes cracked, so there is a limitation as to how aggressively we could hammer the cores down. This leaves open the possibility that we could have penetrated this layer with other equipment, but we were unable to penetrate. It may be bedrock, a boulder layer or something else that prevented us from coring deeper.
   
   
   
   We will add this information to the description of the field work. Here, we will also include that we did not do a core transect for lake stratigraphy and conclusions consequently are open to other interpretations.
   
   
   
   
5. Reporting evidence on six lakes when tsunami layers are only found in two lakes.
   
   Agreed. The revised abstract and manuscript will throughout reflect that we found tsunami deposits in two lakes and used another two to constrain run-up magnitude and timing. The two cores that do not have relevant information, being too short, will be mentioned briefly but not presented in detail.

More author comments (AC) to major comments from the reviewer (RC3) below. Author response to minor comments and suggestions from the reviewer can be found in the file attachment.

Wagner, B., Bennike, O., Klug, M., and Cremer, H.: First indication of Storegga tsunami deposits from East Greenland. J. of Quat. Sci. 22, 321-325. https://doi.org/10.1002/jqs.1064, 2006.

RC3: Fieldwork:

The manuscript presents only one core from each lake basin. Tsunami deposits change a lot over short distances, particularly in a lake basin. To document lateral changes the authors need to map the deposits across the lake basins. Ideally, in order to conclude that this is a tsunami deposit they should have shown that the sediments really were deposited into the lake from the seaside in terms of lateral changes in grain size, thickness and erosion. With only one core this is impossible. There are plenty of examples of lake basins studies using multiple cores, also in Greenland, as done for instance by Long et al., 1999 – a beautiful paper in that respect is that by Long et al., 2015. To document tsunami deposits preserved in a lake stratigraphy you need many cores.

AC: The field work had been planned to take multiple, overlapping cores in each lake, but problems with the core tubes limited this to one per lake. (Two cores were obtained from lake SAQ21-09 and is available for download with the other core data in the data repository. It is not presented in the manuscript, as it contains the same information).

In the fieldwork section and results interpretation in the revised manuscript we will explain that we did not do a core transect and make sure the reader understands the implications thereof.

RC3: The only core is, since it is the only one core, too thin – 60 mm in diameter. This diameter is sometimes too small to find reliable material for radiocarbon dating. A thicker core, or instead several cores of the same unit, would yield a larger volume for each cm core depth to search for terrestrial plant remains for radiocarbon dating.

AC: We agree that multiple cores (or core tubes with larger diameter) would have provided more sediment volume that would in turn have increased the chances of finding dateable terrestrial plant remains. The prerequisite is that there are terrestrial macrofossils present in the sediment - if there is no terrestrial material in the layers, then greater sample volume does not help.

Below is a brief, and certainly not complete, survey of existing literature which shows some of the challenges with finding the optimal terrestrial macrofossil for dating in Greenland.

• Long et al, (1999) and Long et al (2015) exclusively rely on bulk sediment sampling (using 50 mm piston corer taking the part of a core home for lab work, and 40 mm Russian corer for lake stratigraphy).
• Bennike (2000) uses 15 terrestrial plant fossils, 5 limnic plant fossil, and 3 bulk sediment samples to create age models for four lakes. (60 mm piston corer for lab samples, reconnaissance using 50 mm Russian corer).
• Larsen et al. (2017) uses 8 bulk sediment, 6 terrestrial macrofossil, and 1 mixed terrestrial/aquatic mixed for creating age models for three lakes (single core per lake, probably 60 mm piston or percussion).
• Philipps et al. (2017) uses 15 aquatic macrofossil and 1 bulk sediment sample to create age-depth models for two of three lakes in the central Uummannaq Fjord system (the latter dated the contact). (Single core per lake, coring diameter not specified). Core lengths 82, 109 and 113 cm.

In general, the work is done with a single core per lake taken home for lab work, and dependent on the type of investigation, lake stratigraphy or reconnaissance is also done with a Russian corer.

Use of bulk sediment and aquatic macrofossil is widespread and the problems well-known. In the literature above these problems are not introduced to the reader. The reason may have been that since researchers are aware of these problems, and they know the implications are when they see a table of radiocarbon dates where bulk sediment or aquatic samples are present.

Larsen et al. (2022) captures this in the discussion of the use of bulk sediment dating:

“To test if ages of bulk sediment samples are reliable, we compared ages of macrofossils and bulk samples from two intervals in Smaragd Sø and found discrepancies of several centuries (Table 1). Though these are clearly significant, we have used these bulk ages in the study out of necessity. They should conservatively be regarded as the maximum limiting ages for those deposits.”

The problems with aquatic macrofossil and bulk sediment are well-known, but if nothing else is found for dating, then with its limitations it can be used out of necessity. Thus, use of non-optimal dating material is frequent, and in some cases comprises all dates in the study (e.g., Long et al., 1999, 2015).

We have revised our interpretation of our dates elsewhere and refer to this for further details.

Bennike, O: Palaeoecological studies of Holocene lake sediments from west Greenland, Palaeogeography, Palaeoclimatology, Palaeoecology 155, 285–304, 2000.

Larsen, N.K., Astrid Strunk, Laura B. Levy, Jesper Olsen, Anders Bjørk, Torben L. Lauridsen, Erik Jeppesen, Thomas A. Davidson: Strong altitudinal control on the response of local glaciers to Holocene climate change in southwest Greenland, Quaternary Science Reviews 168, 2017.

Larsen, N.K., Siggaard-Andersen, M.L., Bjørk, A.A., Kjeldsen, K.K., Ruter, A., Korsgaard, N.J. & Kjær, K.H.: Holocene ice margin variations of the Greenland Ice Sheet and local glaciers around Sermilik Fjord, southeast Greenland. Quat. Int., 2021. doi.org/10.1016/j.quaint.2021.06.001, 2022.

Long, A.J., Roberts, D.H., and Wright, M.R.: Isolation basin stratigraphy and Holocene relative sea-level change on Arveprinsen Ejland, Disko Bugt, West Greenland. J. Quat. Sci. 14 (4) 323–345, doi.org/10.1002/(SICI)1099-1417(199907)14:4%3C323::AID-JQS442%3E3.0.CO;2-0, 1999.

Long, A.J., Szczuciński, W., and Lawrence. T.: Sedimentary evidence for a mid-Holocene iceberg-generated tsunami in a coastal lake, west Greenland. Arktos 1:6, https://doi.org/10.1007/s41063-015-0007-7, 2015.

Philipps, W., Jason P. Briner, Ole Bennike, Avriel Schweinsberg, Casey Beel, Nathaniel Lifton: Earliest Holocene deglaciation of the central Uummannaq Fjord system, West Greenland, Boreas, 2017. https://doi.org/10.1111/bor.12270

RC3: The cores are short and does not penetrate the entire stratigraphy. The authors say that “All core lengths reflect the thickness of the sediment package until the corer encountered bedrock or impenetrable substrate”. According to the sea level curve and deglaciation dates the lakes should contain marine deposits in their lower parts. The longest cores reach back to 8000 years BP. Also, I wonder, are the top sediments present in the cores? The two cores showing tsunami facies has radiocarbon ages of 5000-6000 years only 20-30 cm below the core top. Has nothing or very little of sediments been deposited in these two lakes since the mid-Holocene?

AC: We will add to the revised manuscript that it is a possibility that there are more sediments in the lakes than what we are presenting, so that this possibility remains open. We cannot know what prevented us from coring deeper and it is a speculative guess.

Lack of top sediment. The sediment dates are indeed quite old close to the top in cores 06 and 09. The best explanation is that we misjudged when we hit the lake bottom and opened the piston too late. We do have the complete core from lake 07 which has a complete top with dates and no trace of tsunami in it.

RC3: Only two of the six lake basins have evidence of tsunami deposits. But the authors write in the abstract: “Here we report evidence from six coastal lakes …..evidence of at least two tsunami events.” One of the six lakes has only a 9 cm core, another has 54 cm long core with a radiocarbon age of 2.8 ka BP. These two basins do not contain any sediments relevant for this study and should not be used to increase the number of sites. It doesn’t look good…

AC: Agreed. The revised abstract and manuscript will throughout reflect that we found tsunami deposits in two lakes and used another two to constrain run-up magnitude and timing. The two cores that do not have relevant information will be mentioned briefly but not presented in detail.

RC3: Radiocarbon ages:

The study has not used the best material for radiocarbon dating. Both bulk organic sediments and aquatic moss is not material recommended for reliable radiocarbon ages. Several studies have shown that aquatic bryophytes yield too old ages because the plants take up their CO2 from dissolved inorganic carbon from the lake water. Even lakes without carbonate rocks in their drainage area could have a reservoir effect – that also could vary quite a lot through time (see many publications about this problem, for instance Marty and Myrbo (2014)).

The other material dated in this study is bulk gyttja. The same problem could arise here, with a lake reservoir age, that sometimes yield too old ages, possibly several hundred years older than terrestrial plant fragments.

AC: Bondevik et al., (1997b) found that radiocarbon dates sampled from just above the tsunami facies commonly show older dates than samples from inside the tsunami facies. The redeposition of older sediments and carbon can occur hundreds of years after the tsunami facies was deposited. This has implications specifically for the bulk sediment sample just above the T1 tsunami facies in core SAQ21-09, which has an age 6.7 cal. ka BP. This date would be significantly affected by redeposition of organic carbon post tsunami and show an age that is significantly older than the time it was deposited. In core SAQ21-06 the youngest date in the core is sampled inside the T1 tsunami facies (as shown in Fig. 4). It is an aquatic moss and provides a maximum age of 5.8 cal. ka BP for the most recent tsunami deposit in the T1 unit, i.e., if the unit represents one landslide-tsunami event then the whole the unit would have this maximum age. If the T1 unit represents multiple landslide-tsunami events, then it would show the maximum age of the youngest event.

Reviewer De Batist (RC2) suggests that a third (i.e., younger) tsunami could explain the young age in top of T1 but could not reconcile this with the fact that there is no trace of tsunami in core SAQ21-07. We also used the bottom date of 7.3 cal. ka BP of core SAQ21-07 to constrain the timing of T1, exactly because there is no trace of tsunami, and it is very unlikely that a tsunami would leave no trace.

Reviewer Bondevik (RC3) points out that aquatic moss when used for dating may show an older age from reservoir effects as the moss could have sourced its carbon from other old organic matter in the water column or if there is infrequent mixing of the lake (Marty & Myrbo, 2014). We do not know the exact species of moss, but it is likely that since the moss obtains its CO2 from the lake water and not the atmosphere, that it could an age considerably larger than the actual age of the deposit. An age of the bottom of core SAQ21-07 that is 1.5 ka too old would explain why there is no trace of a younger tsunami.

This interpretation of the age control makes the 14C dates and facies consistent across the cores.

Our interpretation is that the T1 unit represents at least one tsunami event which occurred in a time bracket of 5.8 and 7.2 cal. ka BP. Reviewer Bondevik thinks that it is possible that T1 represents the tsunami layers found by Long et al. (2015) in southern Disko Bugt. These tsunami layers date to 6.0 cal. ka BP and interestingly fall inside that bracket, so it is a distinct possibility which we will include in the discussion of the Long et al. (2015) paper.

Bondevik, S., Svendsen, J. I., Johnsen, G., Mangerud, J. & Kaland, P. E.: The Storegga tsunami along the Norwegian coast, its age and runup. Boreas, VoI. 26, pp. 29-53. OSIO. ISSN 0300-9483, 1997b.

Marty, J. and Myrbo, A.: Radiocarbon dating suitability of aquatic plant macrofossils, J Paleolimnol, 52:435–443, DOI 10.1007/s10933-014-9796-0, 2014.

RC3: Other comments:

The reader should be presented with a detailed map of each lake including the drainage area together with the location of the core. 

AC: We would prefer not to include an additional figure with detailed maps as we think it would add little to the study. Contextual location of coring sites and sills is presented in lesser detail in Fig 2. Table coordinates for coring sites coordinates lake dimensions, lake depth at coring site and sill heights is shown in Table 2, which also shows overall lake dimensions such as length, width, and area.

We will add a kmz/kml file to the data repository with coring site and sill coordinates, allowing the user to inspect the locations in Google Earth by a double click on the file, or open it in a GIS software.

RC3: What is the age of the T1 tsunami layer? The original interpretation of the tsunami unit T1 includes sediments called “Organic detritus with clasts” (Fig. 4). However, a bryophyte age within these sediments gives an age of 5.8 ka yr BP. This means that the T1 layer is as young or younger than 5.8 ka BP. The authors claim that the T1 layer should be present in the stratigraphy in lake basin 19 m elevation, that is dated to begin at 7300 yr BP, but it is not. The way the authors solve this discrepancy is to place the upper boundary of the tsunami layer T1 at the top of the organic conglomerate layer, and not include the layer of organic detritus with clasts above with the 5.8 ka yr date (Fig. 5). How should we interpret the clasts in the organic detritus layer if not deposited by a tsunami? Could we exclude a tsunami deposit in a lake basin stratigraphy based on ONE SINGLE CORE? I think it is possible that the T1 layer represents the same tsunami event as Long et al., 2015 discovered (around 6.0 ka BP) in the southern part of Disko Bugt. 

AC: Agreed, the 5.8 cal. ka date is in the tsunami facies. The discrepancy between the blue marking of tsunami facies in Figs. 4 & 5 will be fixed so that Fig. 5 will be changed according to the markings in Fig. 4.  This means that the organic detritus with clasts in Fig 5. will be marked as tsunami facies as it is in Fig. 4. Then it will also be consistent with the text in the manuscript.

See also how the interpretation of the radiocarbon ages has been revised in previous response to your comments on radiocarbon ages and its implication on how core SAQ21-07 was used prior to revision to constrain T1 to an age of 7.3 cal ka BP. T1 is now assigned a bracket of 5.8-7.2 cal. ka BP.

RC3: Do the sediments really show two separate tsunami events, or could it be only one event? In core SAQ21-09 the T1 tsunami layer is resting directly on the T2 layer. However, in SAQ21-06 there is about 40 cm of laminated gyttja in between. But is this really original lake sediments and not a rip-up clasts? This problem could have been solved by several cores mapping the layers better in the field. I think it is quite possible that T1 and T2 could reflect one tsunami event.

AC: Figs. 4 & 5 show that in the core SAQ21-06 there is c. 42 cm of laminated gyttja between tsunami units T1 and T2. We are convinced from this that we have at least to distinct tsunami events in the two lakes with sediments and this is based on that the 42 cm of laminated gyttja. Erosion at the base of the tsunami layer seem to be small and radiocarbon dates below, inside, and above the tsunami layer are close.

We will change line 232 to make the possibility that it is a rip-up clast clear to the reader (mainly last sentence):

“We interpret the deposits in the two cores as tsunami deposits. While identification of the tsunami deposits is based on visual description of sediments and structures, and sedimentological proxies, the correlation of the units between the two cores with tsunami deposits is primarily based on the laminated gyttja separating the two tsunami units T1 and T2 in core SAQ21-06, constrained by the age control. This correlation is supported by the rough match of visual appearance and sedimentological proxies of T1 and T2 in the two cores. Since we did not map the entire lake stratigraphy, we cannot exclude the possibility that the c. 42 cm unit of laminated gyttja could be a large rip-up clast and T1 and T2 may be one tsunami event.”

RC3: The comparison to the basins studied by Long et al., 1999 at Arveprinsen and south in the Disko Bugt (Long et al., 2015) is interesting, but depends on the age(s) being correct. See comments above.

AC: With our revised interpretation of the dates the story has changed and the tsunami layer in the south Disko Bugt (Long et al., 2015) now falls inside the age bracket of the T1 layer at Saqqaq.

RC3: See also detailed comments directly on the attached pdf, both text and figures.

AC: Our responses to your detailed comments are available in the attached pdf file.

RC3: Conclusion:

Should this manuscript be published? I wouldn’t publish it, mainly because of only one core from each lake basin. I strongly advise the authors to return to the lake basins and map the deposits using for instance a Russian peat corer. It is important to know if these sediments were deposited from the seaside since they do not contain any marine fossils. Is it really evidence here for two separate tsunami events? The radiocarbon dates should also be tested, are the aquatic moss ages OK? If yes, then the T1 tsunami layer could be younger or as young as 5.8 ka BP – that changes the story - so it is also important to check out the aquatic moss ages.

AC: We would very much like to revisit the lakes at Saqqaq and elsewhere and get much better data, also more distal (to the landslides) sites, but we have work with the data we got home. We believe that with the caveats mentioned above, they have sufficient quality to allow for interpretation and contextualization.

Specific points summarized in this conclusion is addressed elsewhere.