Ice Core Chronologies from the Antarctic Peninsula: The Palmer, Jurassic, and Rendezvous Age-Scales

Round 1

Reviewer 1 Report

The manuscript (MS) presents time scales for three shallow ice cores from the Antarctic Peninsula. The chronologies are obtained from annual layer counting of high-resolution isotope and impurity records and from volcanic linking to the ‘nearby’ WDC ice core. The seasonality of isotopes and impurities is discussed. The study is convincing, the presentation is straightforward and well illustrated, and I only have minor comments.

Line 111: The delta-18-O formula needs editing.

Line 127: Which age markers are referred to here? Annual or volcanic markers? Maybe the seasonality discussion should come after the introduction of annual layers?

Figure 3: This is a nice figure of the dataset from the shallow part of the Palmer core. Could you maybe provide another figure from the deeper section of that core to give us an Impression of what the annual layer counting looks like further back in time?

Figure 4:

- Please indicate the position of Pinatubo and other well-known eruptions in all cores, so that we know what the sulfate profile looks like at the time of the eruption (even if there is no signal). I see in Table 2 that Pinatubo has a defined depth in all of your cores. You can use a different color coding if you want to distinguish between ‘detected’ and ‘non-detected’ events.

- In the caption there is mentioning of a dashed red line. I don’t see it in the figure.

- I’m a bit puzzled about the units of the y-axis on this figure. Normally fluxes of sulfate are in units of ‘kg per km2 per year’ or ‘g per m2 per year’ or similar. With the ‘ppb’ unit you probably refer to mass of nnsSO4 per mass of water? I think you need to get rid of the water in that equation. Also, you may need to introduce a factor of 0.917 to make up for the density difference between ice and melt water. Since you are showing the sulfate record at annual resolution it is implicit that the flux is per year and you need not to include it in the unit. Maybe check out [Plummer et al., 2012] formula (2) for inspiration. Let alone that the records are shown on a log scale so that in principle the units should also be log(units)?

Figure 5: I would suggest to have an overlap of the x-axes in the two subplots: Top: 1900-2015 and Bottom: 1800-1910. You are discussing an event at 1912 and it is difficult to put into context with the present cut at 1909/1910.

Would it make sense to include the WDC nnsSulfur record in one of your figures (Fig 4 or 5) for comparison? There could be other common features (or major differences) than the major volcanic eruptions already discussed?

Line 248-249: Please provide a reference for that statement.

Reference:

Plummer, C. T., M. A. J. Curran, T. D. Van Ommen, S. O. Rasmussen, A. D. Moy, T. R. Vance, H. B. Clausen, B. M. Vinther, and P. A. Mayewski (2012), An independently dated 2000-yr volcanic record from Law Dome, East Antarctica, including a new perspective on the dating of the 1450s CE eruption of Kuwae, Vanuatu, Clim. Past., 8(6), 1929-1940.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors present depth-age chronologies for three shallow ice(firn?) cores drilled across the Antarctic Peninsula region of Antarctica (approx 230 km apart from each other). These cores are all less than ~140 meters in total depth and cover no more than 390 years of total time. From the onset, it is unclear if the primary goal of this paper is to:

- Simply present depth-age scales for three AP cores and discuss how they are constrained to volcanic tie points…and therefore serve as a scale usable for future investigators? If this is the case, nowhere is it noted why these cores were drilled and as a part of what larger project?

- Investigate and compare seasonal deposition of chemical species, and overall accumulation rates, across three cores from a larger (regional) area?

- To simply show validity of annual layer counting methods and speak to the accuracy of their methods when compared to known bipolar volcanic events (mostly as seen in WDC core).

- To stress the importance of measuring multiple cores as it helps validate depth-age scales in places where signals can be drowned out by other sources (e.g. biogenic, or elevation effects, etc).

It is absolutely critical that ice cores be dated properly so that any/all future investigators and studies have a viable scale for which to analyze their data. So with that said, it is important that data such as these are made available. This particular study does lack any sort of testable hypothesis however. That’s not to say that this it should therefore be rejected, but it does put an additional burden on the authors to be as specific as possible with regards to all methods, analyses, calculations, and site properties/characteristics.

It is not clear if this is the first paper submitted with respect to this project and cores (I am assuming it is). If so, I believe it is necessary to discuss in much further detail the site characteristics for each drill location. Table 1 does have lat/long, elevation, drill depths and calculated ages,  but I think it’s critical to also discuss average accumulation rates, temperatures, and firn depth (bubble close-off depths) for each site. These all have major implications with respect to not just the physical properties, but for various chemical and isotopic measurements that could be made (eg. d18O is affected by elevation). As the authors note, accumulation rates are the key indicator for measurable resolution in firn/ice cores, so this piece of information for each site is necessary. What’s more, is due to the shallow nature of these sites, I strongly believe that showing a reconstructed accumulation history for each site is necessary. The reconstruction could certainly be plotted for all three sites on one figure….and this would also allow the authors to further discuss how the thinning corrections were made. (I know that Nye 63 was cited), but perhaps a little more detail on the correction is necessary. In addition, because these cores are primarily firn (at least that’s what’s assumed since no pore-close-off depths are noted), then a density correction would have to be made to normalize annual layer thicknesses to either ice or water equivalent. With respect to site locations, nothing is said with respect to flow characteristics for each site. Is each drilling location on a dome or divide or is there some flank flow (or potential shear)? Could accumulation that’s being measured have been advected from somewhere else (I.e. what is the accumulation’s provenance? is is local). Very importantly…the AP is not a dry firn area and melt does occur. This must have a measurable effect on various isotopic and chemical indicators and should be discussed in the text. Were any melt layers detected in the cores? If so, what could that have done to some of the chemical measurements if melt-water percolated and refroze at alternate depths? how could that affect uncertainties. Lastly, was visual layer counting ever done or considered. If these cores were only 140 meters long, did anyone simply do visual inspection and counting of annual layers as an additional measure of ages? Were visual logs taken in the field or were the cores every viewed on a light table in a dark booth to assess visual stratigraphy?

84 - It would be incredibly helpful here to simply discuss the overarching scientific motivation for the larger project related to these cores. Why were these sites chosen, what hypotheses were being proposed to be tested, and what future analyses are still being done on or with these cores. In short, why were these cores drilled…and as a part of what project?

86 - what defines high-elevation in this context and why is it important to note this? I assume it is because it related to the regional accumulation rates or because it can affect certain indicators (like d18O?), or because it relates to the site's potential to be affected by biogenic sources?

Figure 1 - First off, this figure should have a scale bar for distances. It might also be helpful to shift the regional map slightly to the grid north-east, so that WDC is not hiding all the way in the bottom corner. It may also be useful to simply color WDC a different color to differentiate it from your study sites. I also always recommend running a draft manuscript through an app like "Color Oracle" to check against known color blindness traits for readers. Some colors can be very difficult to differentiate, especially for those with R/G blindness or deficiencies

101 - It would be helpful to note the resolution of the CFA measurements? is it 1 millimeter? 10 mm? Was H2O2 the only indicator measured continuously at all three sites? And why were water isotopes only measured for the Palmer core and not the other two? 

102 - For the  discrete sampling, did this happen on all three cores, and why was 5 cm chosen as the interval? If samples were chosen at 5cm intervals….how thick were the actual samples themselves? The authors mention that the “melted samples were analyzed”, but don’t note how the discrete samples were melted.

107 - why are sodium data missing for the deep part of jurassic core?

111 - It would be helpful for any equations like here (or in line 45 for example) to be put on their own line…regardless of their simplicity.

114 - Again it’s a little confusing in here regarding measurements. It sounds like the Palmer core d18O was measured continuously, but the Rendezvous and Jurassic core d18O measurements were made from discrete 5cm samples? I know situations often dictate what instruments are available and when, but is there a reason the three cores weren’t measured the same? It would reduce multi-instrument uncertainties or the need to integrate measurements over 5cm depth intervals.

125 - It would be useful to specifically explain HOW each year was picked. were annuals picked,  or were they interpolated. (and if both, explain this specifically). It’s unclear in the text if the 5 cm sampling resolution was enough to capture all of the annuals or if some had to be interpolated (something that could be clarified by showing/discussing accumulation records). Given the shallow depths of the cores, it’s assumed that 5cm is less than 1 year in the cores? But again, this is not known as avg accumulation rates are never mentioned.

126 - I’m unclear and a bit concerned about this method and the uncertainties therein. I understand that age markers are picked (although it’s a bit unclear which age markers)…and then dates are linearly interpolated, but wouldn’t this make large assumptions about accumulation rates being constant, no? Is this done because the discrete sampling at 5cm intervals precludes just picking annual layers exactly? With this type of interpolation, wouldn’t that increase overall errors? For monthly means, are the linearly interpolated annual brackets simply broken into 12 equal parts? I understand that the authors are not claiming monthly resolution (which would be unreasonable considering the interpolations), but they are still speaking to the seasonality using monthly median values.

136 - Again, it would be helpful to note specifically how each year was picked. were annuals picked or interpolated.

138 - how were outliers identified?

142 - the authors note the accumulation record here. This should absolutely be shown as a figure in the paper and discussed. It also looks like they note the record is in m of water equivalent, so it should also be noted the calculation/corrections made for that conversion (as well as details on the Nye correction made)

153 - It’s not entirely clear why the 201 moving average was chosen and why 31 years is significant?

169 - would be useful to specify what records were not evaluated and why they did not show clear cycles.

Seasonality and figure 2 - Again, what’s unclear in these plots is which of these data sets were CFA measurements, and which were interpolated from discrete samples? Wouldn’t that have large implications on how the data represented and inherent larger uncertainties for some data sets (especially if the authors are trying to break things down into month-scales)?  5 cm discrete sampling could miss an entire year couldn’t it? How do the authors know if years are missed?  This leads to a follow up question regarding tie points to volcanic events. If the authors note a discrepancy in the dates calculated from the chemical/isotopic data with the volcanic tie points, how is that specifically corrected? Are years simply added in or subtracted out so that the ages match? If that’s the case, wouldn’t that have large impacts on the seasonal/monthly interpolations if years are just added in or subtracted out?

For inset figures, consider maybe swapping month numbers with letters (J F M A M J J A S O N D), or at least be clear in specifying that 1 = january.

183 - authors note that the nssSO4 was the primary chemical for annual layer markers….but weren’t those data measured as 5 cm discrete samples? So again, it’s unclear how how interpolation was factored in. If the answer is simply that due to high accumulation, a single 5 cm sample is much less than a single year…then it just needs to be made clear.

197 - It’s not uncommon for some reconstructions to contain a mix-and-match of different records, but it’s definitely worth spending some time explaining why the range of sampling techniques and instruments were used for these three cores and the resultant implications.

Figure 3 - Were the nssSO4 annuals picked by eye, or was an algorithm used. In addition, were the annuals picked by multiple observers so that proper reproducibility errors could be calculated? It’s cleaer that the nssSO4 has relatively good peaks, but for the depths where nssSo4 is not available, other indicators don’t appear to have as good peak matches. How was this accounted for?

Might be useful to reverse x-axis so that shallower is on the right. This would then equate to more-recent being on right (which is consistent with other figures and fairly standard within the ice-coring community)

210 - This sentence reads a bit odd, and might benefit from a quick reorganziation/simplification.

Figure 4 - figure c shows a dashed yellow line not red (as noted in caption).  I understand picking the larger bipolar events as ones to use as tie points, but am curious as to why the Agung and Pinatubo don’t appear in Palmer, or why say, Krakatoa only appears in Palmer…especially considering the drill sites are less than 250 km apart (Is this due to the overwhelming of biogenic sources?). Also, it would be useful to add small tick marks at 10-year intervals on the x-axis to better identify dates visually.

226 - since the authors note only one significant event for these cores (which obviously makes volcanic tie-ins difficult), couldn’t other eruptions be used that aren’t necessarily the big Bipolar ones? In the Sigl paper they note of other eruptions (e.g. Tarawera or St Maria) that were visible in the WDC core. It does almost appear that there are to sulfate peaks in the rendezvous core at ~ 1895 and 1912 that do push above your 2 sigma line (as well as a few others in the other cores too)

227 - isn’t this statement directly contradicted by your figure? figure 4 b shows several peaks that go above 2 sigma no? or are do the authors mean only 1 event, that is also a bipolar event?

236 - it seems that this point is worth discussing further since it has such large implications on the nssso4 interpretations (and thus volcanic makers). Also, the authors note the accumulation here, but again don’t specify any values for comparison. having read up to this point, I have no idea what the avg accumluation rate is for any of these sites (1 cm a year? 30 cm year? 100 cm year), and therefore no sense of firn depth, or how many annual layers are present per meter.

248 - if this statement is true, doesn’t that mean that stating a 6 month error on your annual dating doesn’t properly account for all uncertainties? Wouldn’t there have to be at least a +- 2 year uncertainty?

249 - along these lines, the authors note a large date different between palmer and wais….but how do they know that their counts/interpolations simply aren’t off by 2 years and that they are an exact match?

251 - probably worth noting that the SEAT WAIS cores were drilled as part of a separate traverse and not related to WDC since this is the first time they are mentioned.

259 - This goes to my previous comment about using additional volcanic markers that perhaps are specific only to the Antarctic. What do the authors think regarding these additional horizons, can they be tied to other cores, and how they could be used to better constrain the depth-age scales?

figure 5 - this is probably the most illustrative/useful with regards to the volcanic picks. Stylistically, it would be useful to use either all solid lines, or all dashed. In addition, it might be improved by putting the older plot (age-wise) on top.

305 - chronologies were evaluated against known volcanic horizons….but it should be made clear if that means they were just compared, or were corrections made in order to actually tie them in and match the time scales (by adding or subtracting years where possible).

316 - why note this? is it because higher-elevation means less accumulation or less influenced by coastal/biogenic sources? 

Citation 18 looks to be published in 2016  not 2015 (just something I happened to notice randomly) - https://journals.ametsoc.org/view/journals/clim/29/7/jcli-d-15-0354.1.xml

Citation 10 notes a doubling in accumulation since 1850. It would seem that this could have implications any accumulation record assumptions made for this study…and further justify plotting the accumulation records for each site.

Other papers that might be worth citing or looking into: 

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2011JD015894

https://onlinelibrary.wiley.com/doi/abs/10.1034/j.1600-0889.1992.t01-2-00007.x

https://www.cambridge.org/core/journals/annals-of-glaciology/article/stableisotopeairtemperature-relationships-in-ice-cores-from-dolleman-island-and-the-palmer-land-plateau-antarctic-peninsula/0BAF9D6060C8A77006BBFC1A78751789

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I am satisfied with the authors' corrections, clarifications and edits to the manuscript. I appreciated their detailed explanations where necessary.