Interactive comment on “Thermal legacy of a large paleolake in Taylor Valley, East Antarctica as evidenced by an airborne electromagnetic survey”

Abstract. Previous studies of the lakes of the McMurdo Dry Valleys have
attempted to constrain lake level history, and results suggest the lakes
have undergone hundreds of meters of lake level change within the last
20 000 years. Past studies have utilized the interpretation of geologic
deposits, lake chemistry, and ice sheet history to deduce lake level
history; however a substantial amount of disagreement remains between the
findings, indicating a need for further investigation using new techniques.
This study utilizes a regional airborne resistivity survey to provide novel
insight into the paleohydrology of the region. Mean resistivity maps
revealed an extensive brine beneath the Lake Fryxell basin, which is
interpreted as a legacy groundwater signal from higher lake levels in the
past. Resistivity data suggest that active permafrost formation has been
ongoing since the onset of lake drainage and that as recently as 1500–4000 years BP, lake levels were over 60 m higher than present. This coincides
with a warmer-than-modern paleoclimate throughout the Holocene inferred by
the nearby Taylor Dome ice core record. Our results indicate Mid to Late
Holocene lake level high stands, which runs counter to previous research
finding a colder and drier era with little hydrologic activity throughout
the last 5000 years.


the Heroic Age. None of the existing datasets is complete. Landforms (mostly deltas) dated by radiocarbon tell you when the lake was at a certain level, but geomorphic records are fragmentary and incomplete. Lake bottom sediments from beneath the present lake can give more continuous records, but their interpretation in terms of specific water level is complicated. Here, the authors give a third way of reconstructing former lake levels, which I'm sure has its own associated problems, but which presents a third perspective that may help us circle ever closer to the actual lake history.
That said, I cannot judge the actual methods used in the remote imaging or in interpretation of the resistivity data. It is outside of my field of expertise. I am assuming here that it is correct, and more reliance should be placed on other reviewers on this aspect.
Although my overall impression of the paper is quite favorable, I do think there are some areas that could use reframing and a critical assumption or two that should be revisited.
Reviewer Comment (RC) 1) The analysis seems to operate on the assumption that there was a continuous drop in lake level from the 80 m sill level to the present lake over the second half of the Holocene (and indeed, perhaps from the highstand of GLW). However, this is unlikely to have been the case. This would have been a closedbasin lake and highly sensitive to changes in inputs (almost entirely glacial meltwater) and outputs (evaporation/sublimation). This hydrology places it among a category of lakes that typically shows outsized reactions to small changes in water balance. Pre-Holocene (and even Holocene) lakes in the Dry Valleys region are suggested to have undergone large-scale fluctuations during the last glaciation and termination of the ice age (i.e., Hall et al., 2010, PNAS -a reference which probably should be included). Lake Vanda showed large fluctuations at 3 and 6 ka. So, it is likely that Lake Fryxell did as well -and possibly other changes not yet documented. To what degree does this lake level volatility affect your modelling results and error bars? How long must permafrost be covered up to completely melt under a lake (I assume this results from permafrost thickness and length of lake cover)? In short, it seems as if this assumption TCD Interactive comment Printer-friendly version Discussion paper may be quite critical to the outcome of the model. Author Comment (AC) 1) We agree that multiple lake level fluctuations in both the filling and draining directions could have occurred throughout the Holocene. Our results do not necessarily assume continuous lake level drop -it would be the same result if the lake level drop was instantaneous. We have made this clearer in the text.
When it comes to analyzing the potential importance of putative lake level variability at elevations <81 m.a.s.l., it is important to note in our Figure 9 that the cluster of thin permafrost at low elevations occurs for elevations <50 m.a.s.l. where permafrost is estimated to be <50 m thick. Comparatively few of our estimates of permafrost thickness fall between 50 and 140 m with corresponding surface elevations between 50 and 81 m.a.s.l. The only other high-abundance cluster of permafrost thickness is 140-200 m with corresponding surface elevations between 81 and 150 m.a.s.l. This cluster represents the lake levels that occurred when the ice dam was still present. When the ice dam is not present, the hydrology of Fryxell basin favors lake levels that are considerably lower than the sill level of 81 m.a.s.l. and that reach mostly up to ca. 50 m.a.s.l., which corresponds to the other cluster of common permafrost thicknesses that are <50m thick. We created a model to explore this effect (fill drain cycles), however we are not including this in the paper because we do not believe we can resolve any individual lake level fluctuation events. Our modeling indicated that permafrost that is <50 m thin takes only ca. 150-300 years to grow, because permafrost growth rates are initially high. Thin permafrost is also most susceptible to melting away relatively fast when submerged by rising lake levels. Hence, it is not likely that our conclusion that the cluster of thin permafrost is very young can be substantially changed by assuming fluctuation of lake levels. ____________________________________________________________________ RC 2) The paper attempts to compare model results with 14C and OSL data. However, this is mixing of apples and oranges (and may stem from the assumption in point 1 above). Both the 14C and OSL dates refer to the position of the lake at a specific time.

Interactive comment
Printer-friendly version Discussion paper They do not preclude the lake from reaching that same level (or higher) at another, later time. It is highly likely that most if not all of the dated deltas were under water at multiple times, not necessarily reflected by the dates (this is to some degree acknowledged in the discussion of OSL, but the same applies to radiocarbon). Some deltas at critical levels (i.e., the sill level) may have been occupied and active at several times. Thus, in my opinion, you cannot make a direct comparison of the three different dating methods, simply on elevation. These lakes fluctuated many times and there are going to be deposits from earlier times mixed in at the same elevations occupied by later lakes. Recognition/resolution of this assumption may help resolve some other issues raised below.
AC 2) We agree totally that 14C, OSL and our model results are measuring different things. But an important point to make is that we are not presenting our permafrost age model as a precise dating technique. It is merely showing that the most recent occupation of the area by a large lake happened on a time scale more consistent with the younger OSL chronology than the older 14C chronology. We also agree that there would have been fluctuating lake levels during the Holocene. But having said that, we note that the OSL data from Berger et al. (2013), if taken at face value, does suggest a fairly smooth curve of lake level decline over the last ∼13k years, but this is hard to say conclusively without more OSL data. The 14C is indeed, all over the place, but does show a general trend of lake level lowering over time. Likewise, the uncertainty of our model results grows large with time.
We understand the reviewer's concern and feel that since we can't resolve the errors in the various approaches, the best thing to do is make sure there is appropriate discussion of the differences in the different data types, and that we are making broad comparisons. We have modified Figure 14 (which compares all three methods), and instead of shading regions with raw data, 1 sigma, and median line, we have displayed the data similar to Figure 9 as individual model points. We still plan to include the OSL and 14C data for comparison. We have rewritten the paragraph starting on line 236 C4 with this goal. Here it is: "With inputs tuned to maximize permafrost ages, it was still not possible to yield ages for lake occupation comparable to those estimated by 14C ages of delta deposits (Hall and Denton, 2000) (Fig. 14). Radiocarbon dating from Hall and Denton (2000) suggests that lake levels above modern levels occurred between 22 to 5 ka BP, and there are no 14C samples to suggest that levels exceed modern lake elevation after 5 ka BP. OSL dates (Berger et al., 2013) estimate past lake level high stands existed between 12 -5 ka BP, and do not suggest recent lake level high stands. Our permafrost age calculations agree more with the OSL chronology than the 14C (Fig. 14), both have limitations. OSL data records the burial age of the quartz grain, which may be different from the deposition age. Rates of past sediment transport and deposition are not currently known, so this lag in deposition versus burial time is unresolved. Secondly, OSL dates are collected from depth, meaning the dates may not be an accurate representation of the most recent occupation of that delta/lake level elevation (this will also be true for 14C samples collected away from the surface). Several studies have shown a substantial radiocarbon reservoir effect in the MDVs, but modern lake edges and streams have been shown to be mostly-well equilibrated with the modern atmosphere (e.g. Doran et al. 1999;Hendy and Hall 2006). Doran et al, 2014 did note some exceptions. In fact, out of 4 moat microbial mat samples dated (all in < 1m of water), only one was equilibrated with the modern atmosphere. The others carried 14C ages of 2324±96, 9334±71, 2608±48. So clearly, even in modern lakes it is possible for lake edge material to carry a significant carbon reservoir. A large glacial lake of the past may have had even more ancient unequilibrated carbon associated with it due to melt coming from the ancient RIS without the opportunity for significant equalization with the atmosphere (e.g. through direct subaqueous glacial melt and moats being more closed to the atmosphere)." ____________________________________________________________________ RC 3) I am having some difficulty with Fig. 14. If I am understanding it correctly, parts of C5 the modelled curve become problematic and require some explanation. I have already mentioned the 14C and OSL data -I don't think you can compare them directly for reasons mentioned above and recommend taking them off this diagram. My concern here is about the model results themselves.
My understanding is that this figure shows the mean of the model (solid line), a dark shaded zone (1 sigma), and a light shaded zone (raw data). The underlying assumption here is that the permafrost age is directly linked to the lake presence (let's revisit this in a minute). Thus, the older permafrost ages at higher elevations are attributable to the lake dropping from there earlier. Those at >80 m must be from a time when there was still an ice dam at the valley mouth, because they are higher than the sill. However, model ages for >80 m seem far too young. Measurements by multiple methods place deglaciation at the mouth of Taylor Valley by 8 ka if not a bit earlier. On Fig. 14, 8 ka intersects the model results (solid line) at 160 m elevation, something that is not possible, because there would have been no ice to dam the lake. Even the 1 sigma error bars are well above the 80 m sill. The lake cannot exist above 80 m after 8 ka, so how is this result explained? Only the low end of the raw data fit in the plausible zone (>80 m = >8ka). Is there something about the model or in the picking of 100 m that is producing results skewed toward young ages? How does this affect the reliability of the conclusions in this paper? In other words, if we know the model is producing erroneous ages above 80 m, why should we have confidence in results below 80 m elevation? Greater discussion would be useful here.
AC 3) We appreciate this comment from the reviewer, and agree that this topic deserves more discussion. It is correct that the assumption that permafrost age is directly linked to the presence of a lake. However, the exact date that the ice dam is removed is still not completely resolved in the literature. Based on OSL, Berger et al. (2013) suggests a lake level drop below sill level between 5 to 8ka. Hall and Denton (2000) reports one 14C sample from Explorer's Cove that suggests Ross Ice Sheet was last grounded in Explorer's Cove ∼6ka BP. Additionally, Levy et al. (2017) uses lumines-cence dating to suggest that large paleolakes in neighboring Garwood Valley persisted well after the Ross Ice Sheet retreated. The paleolake Howard (Garwood Valley) was at its maximum elevation until 4.26 +/-0.72 ka because of stranded ice in the mouth of Garwood Valley that was a relic of the Ross Ice Sheet. Levy et al. (2017) makes the case that relic ice could have persisted in the mouths of various valleys well after the Ross Ice Sheet retreated in McMurdo Sound. Levy et al. (2017) points out that radiocarbon samples (14C) have consistently produced dates 5 -10 ka older than OSL samples from the same locations, so we think our paper is within this argument that 14C likely overestimates paleolake age by thousands of years. For this reason, we think it is reasonable, based on the literature, to consider younger (than 8 ka) ages for a lake above 81 m asl.
Because of this reviewer's comment as well as another reviewer's comments, we have decided to include a second approach to calculating permafrost ages using a 1D numerical (finite-difference) model solving the classical Stefan problem of vertical heat diffusion coupled with latent heat release during freezing (see attached Figure A). The upper boundary condition is a prescribed temperature that is deltaT lower than the freezing point of the sub-permafrost brines (deltaT = surface temperature -brine freezing temperature). This deltaT can be either held constant during numerical experiments or can be prescribed to vary with time. The bottom boundary condition of the numerical model is a constant heat flux, set to the geothermal flux of 0.080 W/m2 consistent with two borehole-based estimates proximal to the study area (boreholes DVDP-6 and CIROS-1 in table 1, Morin et al., 2010). Other model parameters are based on permafrost properties listed in Table 1 in our original manuscript.
Existing observational constraints indicate that under modern conditions in the study area the temperature at the bottom of the permafrost is ca. -9C (e.g., Figure 7 in Foley et al., 2015) while ground surface temperature is ca. -19C (Table 1 in Obryk et al., 2020), yielding deltaT of about 10C. When we assume a constant deltaT of 10C, our numerical experiment still yields fairly young ages for elevations below the 81 masl sill C7 TCD Interactive comment Printer-friendly version Discussion paper level (<4ka). We then applied a linear cooling rate of 1, 2, 3, and 4C over the last 10,000 yr to model the cooling trend observed in the Holocene Taylor Dome paleotemperature reconstructions (Stieg et al., 2000, Monnin et al., 2004. These ice core constraints are best approximated by a linear cooling trend of 3C per 10,000 years. In the revised version of the manuscript we have included these results to create an additional figure (similar to Figure 13) which shows the permafrost age distribution using the 1D vertical diffusion model. This method still yields fairly young ages for elevations below the 81 masl sill level (<150 m thick permafrost results in ∼4ka, green line below). This would be about 2.5 ka older than our original approach using Eq 1 (Osterkamp and Burns, 2003). Even though the results vary between the two methods, the data still suggests that a lake occupied Fryxell basin between 1.5ka (analytical method) to 4ka (1D vertical diffusion model explained above). Higher elevations (permafrost thicknesses >/=200 m) yields ages between 7 -8ka. This paper discusses Glacial Lake Washburn in the context of past lake level highstands, however we are not trying to necessarily date the timing of the highest lake levels. The errors in our approach are greatest where permafrost is thickest (and hence oldest), so other methods to determine very old lake levels are likely better than our method of using resistivity data. We are simply suggesting that lake levels have been at sill level within the last 5ka. ____________________________________________________________________ RC 4) While it seems plausible, is changing lake level the only variable that can affect permafrost age? How old is that brine?
AC 4) An extensive body of scientific work documents that surface climatic and hydrologic conditions in Taylor Valley have varied considerably on timescales ranging from decades to millennia (e.g., Hall et al., 2000;Doran et al., 2002, etc.) Much less is known about subsurface variability in the study region. However, it is generally C8 TCD Interactive comment Printer-friendly version Discussion paper assumed that on Earth the subsurface is hydrologically and thermally less dynamic than subsurface. The diffusive nature of heat transfer near the surface of the Earth rapidly mutes the amplitude of surface climate fluctuations as depth below the surface increases. For instance, Clow and Waddington (1996) reported a re-measurement of the vertical temperature profile in the DVDP-11 borehole, which is contained within our study area. They observed that 1degreeC of ground surface warming between the 1970s and 1990s decade to nil within ca. 70 meters from the surface. We had the opportunity to re-measure subsurface resistivity in the lower part of Taylor valley between 2011 and 2018. Our comparison of the two datasets does not reveal any discernible subsurface resistivity changes. We also used the regional measurements of resistivity to infer that the subsurface groundwater system in Taylor Valley and neighboring valleys is quite sluggish (Foley et al., 2019). Our upper-end estimate for groundwater velocity between Lake Fryxell and the mouth of the Taylor Valley is 0.01 m/yr. This means that even within the longest timescales considered in our manuscript (ca. 10,000 years), groundwater may have moved horizontally by a total of <100 meters. The only other process we can think of that may, in principle, contribute to temporal variability of properties in the sub-permafrost brine is molecular diffusion of ions in pore spaces. However, the diffusivity coefficients for ions like Na+, Cl-, and other major constituents of Taylor Valley brines is orders of magnitude lower than thermal diffusivity so this process should be even more sluggish than the diffusion of surface temperature changes that we considered above. Therefore, we expect the properties of the sub-permafrost brine to be relatively constant on timescales of centuries and millennia considered in our manuscript. When it comes to the related question of the age of these brines, there are no direct estimates for the subsurface brines in Lower Taylor Valley because these brines have never been sampled. However, the brines in Lake Bonney and at Blood Falls have been studied in more detail and most recent estimates of their age invoke ages of the order of hundreds of thousands of years to millions of years (e.g., Mikucki et a., 2009;2015;Foley et al., 2019;Blackburn 2020). These brines were formed through cryoconcentration of seawater and the last time C9

Interactive comment
Printer-friendly version Discussion paper that seawater may have been emplaced into the subsurface of this coastal region is in the Pliocene or Miocene. This is one reason why our model of the subsurface brine system in Taylor Valley envisions the brines to be old and slowly evolving. RC 7) Line 61 -"atmospheric correction" is an odd way to put it. I would simply call them calibrated ages. Hall et al., 2010 in PNAS is a key paper here with an updated dataset (see the SI with that paper).
AC 7) We agree with the reviewer and the sentence now reads "The calibrated 14C ages from Hall and Denton (2000) shown in Figure  4 were corrected using the CALIB program (Stuiver et al., 2018)." ____________________________________________________________________ C10 RC 8) Line 67 -I'm not sure the lack of samples led to an assumption that lakes remained at or below modern levels over the last 5000 years but only that there were no data and that Lake Fryxell had to be below the sill.
AC 8) We agree that the lack of younger than 5k yr deltas is suggestive, not conclusive, and have removed this sentence. ____________________________________________________________________ RC 9) Line 106 -says <1000 m but line 115 and elsewhere uses 100 m as the cutoff. Is this a typo or something else? If not a typo, why was 100 chosen? AC 9) We chose 100 ohm-m as our boundary between permafrost and brine, so this wasn't a typo. However, the reviewer's comment shows that we need to clarify why we chose 100 ohm-m, and why we mentioned 1000 ohm-m as another threshold.
Permafrost can be broken into subgroups, depending on degree of saturation and confining properties as described in McGinnis and Jensen (1971). Confining permafrost does not allow any fluid flow, whereas aquifrost is permafrost that allows groundwater flow due to local conditions such as salinity and porosity (McGinnis and Jensen, 1971). Confining layer permafrost tends to have much higher electrical resistivities (<10,000 Ωm) than aquifrosts (50 -1,000 Ωm) depending on temperature and degree of saturation of the aquifrost (McGinnis and Jensen, 1971).
The 1000 ohm-m threshold comes from both McGinnis and Jensen (1971) as well as the Mikucki et al. (2015) paper which broadly defines sediments as brine-bearing if they have resistivity from 10 to 1000 ohm-m. We are using the term "high" versus "low" resistivity to broadly explain how the depth of investigation varies.
The 100 ohm-m cut off was chosen using Foley et al. (2015, Figure 7) Dry Valley Drilling Project (DVDP) borehole data comparison that shows resistivity measurements. The borehole DVDP-11 was drilled within our study area and had temperature and salinity measured. These borehole data can be compared to our resistivity pro-C11 file taken near the borehole site (Foley et al., 2015, Figure 7). Their figure shows a rapid decline in resistivity from about 100 ohm-m down to down to <5 ohm-m within only 20-30 m. This sharp resistivity gradient corresponds well to an abrupt jump in salinity with depth. Such a salinity increase is consistent with a transition from frozen sediments to liquid brine. This is how we justify our choice of 100 ohm-m as the nominal boundary between permafrost and liquid brine. If we were to choose a threshold value lower than that, the permafrost thicknesses we would get would be marginally greater (ca. 20-30 m). If we used a higher cut off (say 500 ohm-m), the permafrost layers would be even thinner, and therefore yield even younger ages. ____________________________________________________________________ RC 10) Line 196 -see comment above about these deltas being covered and recovered by water and thus direct comparison is impossible.
AC 10) This will be covered by a new paragraph in the Discussion, and we have added words of warning of direct comparison throughout. ____________________________________________________________________ RC 11) Line 205 -this distance would be made more useful for lake level if an elevation were associated with it.
AC 11) We used a distance from lake edge for this case because the groundwater system is assumed to be confined by permafrost from above and bedrock below. The groundwater system extends laterally from the lake and has been freezing back after lake levels dropped.
Instead of changing the text, we added a reference to Figure 7 at the end of the sentence on Line 205. ____________________________________________________________________ RC 12) Line 213 -this assumption is critical and might be key.
AC 12) It is not clear to us that this assumption is as consequential to our results and their discussion as the reviewer presupposes. One of the most important issues C12 treated in our paper is the question of when the Fryxell basin was filled up with water to above the sill level of ca. 81 m.a.s.l. Such a deep lake required an ice dam. We are focused on using the estimated permafrost thickness to estimate when the ice dam was last present at the mouth of Taylor Valley. Once the ice dam was gone, it did not come back so lake levels could not reach above 81 m.a.s.l anymore. Hence, the issue of lake-level cycling is not really relevant to our analyses of the permafrost with surface elevation >81 m.a.s.l. One could argue that maybe paleo-lake levels varied above 81 m.a.s.l. before the final disappearance of the ice dam. This could result in some of the permafrost thickness observed at these elevations being formed already before the final disappearance of the ice dam. This, in turn, would push our estimate of the final ice dam disappearance to be younger than it already is. See above comment (response to Major Point 1) for further explanation about how < 81 masl (below sill level) fluctuations would impact permafrost age.
In order to better address and highlight these assumptions, we have provided further discussion about how Tps and lake level fluctuation would affect the resistivity signal. Another reviewer also pointed this out as needing further explanation.
Regarding constant Tps assumption, as mentioned above, we have included a second 1D model that allows us to prescribe a variable Tps through time (we chose a linear cooling rate of 3C over 10000 years). This yields slightly older ages.
Regarding the constant lake level drop assumption -we will clarify that this analytical approach does not distinguish between constant drop or instantaneous drop, but either way it does not take into account multiple cycles of fill/drop that could have definitely occurred. This is especially true for recent lake level history without an ice dam (<81 masl). As mentioned above, we applied a cyclical fil/drain cycle to the 1D model. We have not included this figure in the paper, but have provided more explanation about what various fill/drain cycles would mean for our interpretations. ____________________________________________________________________ C13 RC 13) Line 225 -this paragraph goes with the assumption of gradually lowering lake, which was not the case. See Hall et al., 2010. It also makes the assumption of a close link between air temperature at Taylor Dome and lake level, which may be problematic. Also, see comment above about revised Taylor Dome Holocene timescale.
AC 13) Our results do not necessarily assume continuous lake level drop -it would be the same result if the lake level drop was instantaneous. See comments above. We make sure to clarify this in the text.
We will rewrite this paragraph starting on Line 225 to remove the definitive language talking about exact ice dam removal, because as this reviewer pointed out, our results show younger permafrost ages above the sill level.
Regarding the revised Taylor Dome Holocene timescale and associated temperatures, we mentioned above that we will replace the Steig et al. (2000) Holocene temperature ( Figure 2) with a modified figure from Monnin et al. (2004). We understand the reviewer's comment about the link between air temperature at Taylor Dome and lake level, and will give this topic more discussion. We recognize that there is more that goes into lake level change than simply air temperature alone: changes in albedo (snow on landscape) and incoming radiation will change the energy balance, and subsequently the stream/lake response, but you have to have temperatures near freezing for melt to happen. These limitations will also be included in order to clarify the dynamic relationship between lake level and meteorological conditions. ____________________________________________________________________ RC 14) Line 238 -I may have missed it, but I didn't see anywhere in Hall and Denton (2000) where they stated that lake levels had remained at or below present since 5 ka. They didn't have data to address this.
AC 14) We agree and have clarified this assumption throughout the paper (mentioned above in previous comment as well). Specific to Line 238, we rewrote it saying "Radiocarbon dating from Hall and Denton (2000) suggests that lake levels C14 were higher than modern levels between 22 to 5 ka BP, however Hall and Denton (2000) do not provide evidence of lake levels exceeding modern levels after 5ka." ____________________________________________________________________ RC 15) Line 243 -the sentence about OSL dates from depth and not applying to most recent episode of lake level change applies equally to radiocarbon.
AC 15) We agree and have made changes in this section of the paper to reflect this. ____________________________________________________________________ RC 16) Line 249 -application of a large reservoir correction to shallow water lake sediments seems ad hoc, given prior references. Most of these deltas formed at the same stream mouths as today, fed locally by the same glaciers as today, not the RIS. As the data should not be compared directly (for reasons already given), such an attempt at correction is not warranted.
AC 16) See comments on this above. The data we are comparing in this section of the paper is 14C and OSL. The argument we are making is that in a large glacial lake, there may have been more pathways for ancient carbon to remain unequilibrated with the atmosphere -for instance -at the glacier edge there would have been a large subaqueous ice wall melting directly into the lake without contact with the atmosphere. There would have been large volumes of supraglacial melt in contact with and eroding ancient ice until it entered the lake. The moats may have been much less open than today. None of this seems unreasonable, especially when considering moat microbial mat ages in the modern lakes shown by Doran et al 2013 between 2k and 6k 14C yrs. The preserved mat material in the ancient deltas is from these lake edge mats the stream bed is moving over, not the stream beds themselves. ____________________________________________________________________ RC 17) Line 259 Whittaker here and later in paragraph.
AC 17) Thank you for pointing out this missing reference.
We have added the following reference. Whittaker, T., Hall, B., Hendy, C., and Spaulding, S. Holocene depositional environments and surface-level changes at Lake Fryxell, Antarctica.
The Holocene, Volume 18, Issue 5, Pages 775-786 (2008) ____________________________________________________________________ RC 18) Par. starting with 265. While I don't necessarily disagree with the conclusion of the last drop in lake level being post 1.5 ka, I find these stated reasons unconvincing. The link with the Taylor Dome ice core here implies, without specifically stating, that high temperatures should lead to high lake levels. But, this may not be the case. If you want to make this argument, this needs to be stated explicitly (and you'll want to confirm that the chronology didn't shift too much for this to still apply). In addition, there are many, large, well-preserved deltas and while the Crescent Stream one is a nice one, weathering rates are so low that you cannot resolve 1000 yrs vs 10,000 years based on delta appearance.
AC 18) We switched to Monnin et al. (2004) timescales per the recommendation of this reviewer, and their reconstruction suggests that temperatures were above modern from ∼1,800 to 12,000 yr BP. This new record stops at 1ka unfortunately.
Similar to the reviewer's comment about Line 225. We have taken the reviewer's suggestion of explicitly stating the relationship between temperature and lake level. We will support this claim using the following references: Obryk et al. (2017), shows how warming due to foehn winds could be the cause of higher lake levels during the LGM. Doran et al. (2008) which shows how a summer with record single season lake level change was connected to temperature (degree days above freezing). And Doran et al. (2001) which shows a decadal lake level decrease due to cooling. Like mentioned above in the response to Line 225 comment, we recognize that other meteorological variables impact lake level and have highlighted this.
We are not trying to date the delta from its appearance, but simply pointing out that the best preserved delta in the entire Fryxell basin is at the sill elevation.
We have included a sentence to further highlight that we cannot quantify the relationship between delta appearance with delta age due to many unknown variables (age, deposition rates, erosion rates, etc). ____________________________________________________________________ RC 19) Be careful with the 6 degrees warmer comments about Taylor Dome. Temperature is only one of several variables that affect stable isotope data. In addition, the 100 yr running average in the Steig paper (probably more applicable to lake-level changes than any single year) suggests only about 1 degree change and is probably a better representation of actual temperature changes. I think these reasons don't help your argument any. Why not just say your data suggest the lake might have last been at the level of the sill at 1.5-1 ka and leave it at that? To my knowledge, there is nothing that says it couldn't have been. The OSL and 14C data are older, but they may be dating earlier events.

AC 19)
We agree that the 6C should be removed, especially because we took the reviewer's suggestion to use Monnin et al. (2004) updated timeline. We will simply point out that temperatures were warmer than modern during the Holocene. ____________________________________________________________________ RC 20) Line 287 and elsewhere. GLW refers only to the ice-dammed lake. Once the ice sheet retreated, this became Lake Fryxell. AC 21) We will check all references, and as mentioned above, will add Whittaker et al.
(  4) suggest changing "atm corrected" to "calibrated" We agree, and we changed the legend to read "calibrated 14C". ____________________________________________________________________ Fig 6) I found the color on these maps (this and the ones that follow) hard to follow, I'm afraid. Is there a better way to present this? The caption could be more informative for those (like me) who do not deal in resistivity commonly. Perhaps you could add a sentence about what is meant by constant elevation.
This exact rainbow color scale is very commonly used for airborne geophysics, and is close to being the de facto standard. A linear color would produce images where the structure is visible, but the values on the figure would be completely unreadable. Also a linear scale would be extremely difficult to capture the variation across three orders of magnitude. The log scale balances seeing contrasts in both the low and high resistivity limits (which is needed in this region). We believe it would be far more difficult to find a good looking linear scale that can capture the variations in our resistivity maps that the log scale currently shows well. For these reasons, we propose keeping the color scale as is, but will defer to the editor to make the final call or provide guidance.
Thank you for this suggestion about the caption, and we will include the following sentence.
"Constant elevation refers to a horizontal slice through the subsurface at depth (in this case, 100 m below sea level) that is 5m thick." We may also include a light grey shading of this elevation and thickness on Figure 7 to help readers connect the two figures. ____________________________________________________________________ Fig 13) see comment above about using GLW. This should be "lowering of Lake Fryxell" rather than draining of GLW, which would have happened when the ice sheet retreated. Draining really isn't the right word for something that evaporated.
Thank you for pointing this out and we agree it is an important distinction.
We will pay special attention to the usage of the term Glacial Lake Washburn, and reserve it for lake levels higher than the sill level (81 masl). ____________________________________________________________________ Fig 14) was covered in comments above. Suggest removing other chronologic data as the comparison doesn't really mean anything and is potentially misleading.
We addressed the comments about apples and oranges elsewhere. We feel we have provided enough context, caveats and explanation of the various approaches and what "age" means (they all report age) that this comparison is appropriate. The main takeaway is there is some discrepancy between the 14C and OSL chronology, and the permafrost modeling agrees more closely with the latter. It is difficult to reconcile the permafrost ages with the older 14C chronology, no matter how we turn the knobs of the model. We are not trying to establish permafrost age modeling as a dating technique, but in this case it acts as an independent check on the two chronologies.

____________________________________________________________________
Overall Impression Although I have pointed out several potential issues with the paper, I think the approach is really intriguing and the study is important to publish. I think most of these comments can be dealt with fairly quickly. Assessing the impact of the assumption about constant lake-level drop (which is critical), may take more effort. I recommend publication after moderate revision.