the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Late Holocene glacier and climate fluctuations in the Mackenzie and Selwyn Mountain Ranges, Northwest Canada
Adam Christopher Hawkins
Brian Menounos
Brent M. Goehring
Gerald Osborn
Ben M. Pelto
Christopher M. Darvill
Joerg M. Schaefer
Abstract. Over the last century, northwestern Canada experienced some of the highest rates of tropospheric warming globally, which caused glaciers in the region to rapidly retreat. Our study seeks to extend the record of glacier fluctuations and assess climate drivers prior to the instrumental record in the Mackenzie and Selwyn Mountains of northwestern Canada. We collected 27 10Be surface exposure ages across nine cirque and valley glacier moraines to constrain the timing of their emplacement. Cirque and valley glaciers in this region reached their greatest Holocene extents in the latter half of the Little Ice Age (1600–1850 CE). Four erratics, 10–250 m distal from late Holocene moraines, yielded 10Be exposure ages of 10.9–11.6 ka, demonstrating that by ca. 11 ka, alpine glaciers were no more extensive than during the last several hundred years. Estimated temperature change obtained through reconstruction of equilibrium line altitudes show that since ca. 1850 CE, mean annual temperatures rose 0.2–2.3 °C. We use our glacier chronology and the Open Global Glacier Model (OGGM) to estimate that since 850 CE, glaciers in this region reached a maximum total volume of 34–38 km3 between 1765–1855 CE and have lost nearly half their ice volume by 2019 CE. OGGM was unable to produce modeled glacier lengths that match the timing or magnitude of the maximum glacier extent indicated by the 10Be chronology. However, when applied to the entire Mackenzie and Selwyn Mountain region, past-millennium OGGM simulations using the Max Planck Institute Earth System Model (MPI-ESM) and the Community Climate System Model 4 (CCSM4) yield late Holocene glacier volume change temporally consistent with our moraine and remote sensing record, while the Meteorological Research Institute Earth System Model 2 (MRI-ESM2) and the Model for Interdisciplinary Research on Climate (MIROC) fail to produce modeled glacier change consistent with our glacier chronology. Finally, OGGM forced by future climate projections under varying greenhouse gas emissions scenarios predict 85 to over 97 % glacier volume loss by the end of the 21st century. The loss of glaciers from this region will have profound impacts to local ecosystems and communities that rely on meltwaters from glacierized catchments.
- Preprint
(26311 KB) -
Supplement
(186294 KB) - BibTeX
- EndNote
Adam Christopher Hawkins et al.
Status: open (until 20 Jun 2023)
-
RC1: 'Comment on tc-2023-55', Christopher Halsted, 26 May 2023
reply
General Comments
This manuscript represents a valuable contribution to our understanding of Holocene glacier chronologies and regional paleoclimate fluctuations in northwestern Canada, particularly given the relative lack of empirical data from this remote region. The glaciers and moraines targeted by the authors are well-suited for the study objectives, an impressive feat given that surveying was done through satellite and aerial imagery. The methods are generally appropriate, although I have some critiques about how the 10Be exposure ages were statistically interpreted (see following sections). I am not as familiar with ELA reconstructions or climate modeling as I am with exposure dating, but the methods, assumptions, and applications seem reasonable as conducted here. The authors do a good job of comparing their interpretations of Holocene glacier chronologies to other nearby glacier and paleoclimate records, providing a nice synthesis of climate change in the past millennium in northwestern North America.
The foundation of this manuscript is solid, but there is some work that needs to be done organizationally and in terms of data analysis before I can recommend it for publication. I outline my specific comments below. I hope that the authors find these comments to be constructive and helpful, rather than onerous.
Specific Comments
Aside from smaller technical comments, I have two more substantial and specific critiques for the authors to consider.
First, this manuscript does not have a background section, but I believe that it would benefit from one. As written, a lot of background information is sprinkled between the methods, results, and discussion sections, such that the methods section is very long (6 pages) and some much-needed background about the methods being used is introduced after the results have been presented. The existing “Study Area” section, which currently consists of a single paragraph, could also be wrapped into the background. You might also consider adding some field or site photos to this background section, especially because your field area looks stunning (perhaps your SM Figure 7?). I have noted in the “Technical Comments” section the specific lines that I identify as being more background than methods and could be re-located to a background section. Additionally, there is currently limited background about 10Be exposure dating, although it is a key component of this study. Consider expanding the background information about exposure dating, including the issue of inherited nuclides causing age scatter that is so prevalent in glacial moraine chronologies (see Balco, 2020, in Annual Review of Earth and Planetary Sciences for a great overview). As is, inheritance is only mentioned once at the very end of the discussion, but I believe that it plays a substantial role in some of the older exposure ages observed on moraines in this study.
On that note, I have some critiques about how moraine abandonment dates were estimated from 10Be ages. The significant variation in ages on several moraines suggests some source of geologic scatter, rather than just being due to analytical uncertainties, but the potential causes of this scatter are not considered. Rather, all ages are used to estimate moraine ages, causing 1) considerable disagreement between some moraine ages and 2) some very large age uncertainties, especially for moraines with several older exposure ages (e.g., Butterfly, Mordor outer, and North Moraine Hill glaciers). In my opinion, there are two plausible explanations for the observed exposure age scatter, and they bear consideration at some point in the manuscript. First, boulders with older exposure ages (~1 to 4 ka) may contain varying amounts of 10Be inheritance. If your sampled glaciers were indeed less extensive for the majority of the Holocene than during the LIA, these boulders may have been exposed on the proglacial landscape for thousands of years, accumulating 10Be. During the LIA, the boulders would have been re-worked onto the moraines as glaciers advanced, but they may not have been entirely stripped of their Holocene 10Be. If this history is correct, the scatter observed among these older ages likely reflects both re-orientation and varying degrees of glacial erosion experienced by these boulders during LIA re-working. Another potential mechanism to explain the geologic scatter is that the younger ages reflect post-deposition processes that result in partial shielding or disturbance of moraine boulders, thus causing their ages to be younger than the true moraine abandonment date. In this case, the older moraine boulder ages would more accurately reflect the dates of moraine abandonment, and the young ages are ‘red herrings’. In my opinion, the first explanation (older ages have inherited 10Be) is far more plausible, especially given the tight distribution observed in your younger ages across moraines and the contrastingly large distribution of older exposure ages (as demonstrated well in figure 3).
I say the above not to be overly critical, but because I genuinely believe that you have a valuable dataset here and that significant results are being overlooked because of the analyses used. If I may offer a suggestion, I’d recommend labelling the older ages as outliers containing 10Be inheritance and estimating moraine ages using the mean and standard error of the younger ages. Such an approach seems warranted when looking at exposure ages from all sampled moraines together. If we assume that all of these moraines correspond to approximately the same paleoclimate event, and are thus of similar age, then the distribution of exposure ages shown in Figure 3 clearly demonstrates that the young ages are tightly clustered while older ages exhibit quite a lot of variance (likely due to varying levels of 10Be inheritance in sampled boulder surfaces). I think that by using just the young exposure ages, you will get much tighter and more consistent age estimates of moraines, and the overall age estimate of the moraine population will likely become younger as you remove the older samples. You already do this to an extent from lines 285 to 287, where you identify the peak of exposure ages, but I think you can use this peak as evidence to get more accurate moraine ages by discarding old ages.
Technical Corrections
Line 60: The wording of this line is a bit confusing, maybe re-write as “…reached their greatest Holocene positions around 1600-1850 CE, at the culmination of the Little Ice Age (LIA, ~1300-1850 CE)…”
Lines 62-64: Consider moving the last line to the beginning of the paragraph, it reads like a topic sentence (which is missing from this paragraph anyways)
Lines 65-69: The use of a numbering system for only the first two objectives is somewhat confusing. Consider either numbering your third and fourth objectives or get rid of the numbers for the first and second objectives.
Lines 79-82: The climatological information as it is presented here does not feel strictly relevant to your study. It becomes relevant later when you discuss the paleoclimate implications of your ELA reconstructions, but that connection is not clear as written. Consider either adding a few lines explaining its relevance to Holocene glacier chronologies, or remove this information here and bring it up at the relevant point in your discussion.
Line 84: Remove “To summarize our methods”
Line 89: Consider adding something like “…and infer changes in temperature and precipitation from estimated ELA changes” to clarify how you are inferring the paleoclimate changes
Lines 105 – 107: The wording of these sentences is confusing, because you introduce the glaciers by name and then state that most have no formal name. Could you re-word so that it is clearer that those are your informal names for the glaciers?
Line 105: Should this citation be for SM Table 1?
Line 125: SM Figure 3 does not seem like the right figure to be citing here (it is a climate model temp and precip bias calibration)
Line 127: Add a reference to support your statement about moraine boulders (I recommend Heyman et al., 2016, Quaternary Geochronology).
Line 136: Consider changing this sentence to “We processed samples collected in 2014 at the Lamont-Doherty…” As is, it sounds like it was LDEO itself that was doing the sample processing, rather than you.
Line 139: Consider replacing the Nichols and Goehring reference with Kohl and Nishiizumi (1992). The Nichols and Goehring paper was specifically about complications in quartz isolation for in situ 14C exposure dating, which is not relevant to this study. Kohl and Nishiizumi is the original (and still followed) quartz isolation procedure for 10Be analysis.
Line 141: Should this citation be SM Table 3?
Line 144: Add a reference to Table 1 somewhere in this sentence, as Table 1 shows which samples were sent to PRIME vs. LLNL. Also, for consistency, either give the abbreviation for LLNL-CAMS after introducing the laboratory, or don’t give the PRIME abbreviation in the text.
Table 1: Consider rounding exposure ages and uncertainties to the nearest decade. Annual precision is not yet feasible with 10Be exposure dating. Additionally, the caption repeats information given in the table footnotes (about the exclusion of erratic boulders from moraine ages), consider deleting this information in the caption. Finally, grammar edit for footnote d: “excludes the exposure age of erratics, whose ages are listed in italics”
Lines 154-159: All but the last line of this paragraph feels like background information, it should probably not be part of the methods section. See my comments in the “Specific Comments” section about maybe adding a background section, these lines would fit into such a section to guide the reader through using ELAs to reconstruct past climates. Also consider expanding your explanation of “Each method offers advantages and limitations in reconstructing past ELAs”. Some readers (myself included) might not know the systematics of these methods and require a bit more guidance.
Line 156: Remove apostrophe from “ELA’s” here and elsewhere`
Lines 165-169, 171, and 183-186: The first sentences of the THAR, AAR, and ELA/precipitation paragraphs also feel like background information rather than explanations of your methods
Line 250: “Finally” is used in successive paragraphs, consider removing it from this sentence.
Lines 262-279: These two paragraphs also feel like background information that should more appropriately belong in your “Study Area” section, or in a dedicated background section.
Lines 276-281: The ages of the moraines are likely to change if you follow the suggestions I provided in the “Specific Comments” section, but as is, the large uncertainties on exposure ages should probably be mirrored in the dates you give in parentheses. For example, “610 ± 850 a (ca. 1405 CE)” should realistically read “610 ± 850 a (ca. 1405 ± 850 CE)” or “(ca. 560 CE – Present)”
The top panel of Figure 3 is great. However, I wonder if box and whisker plots are the best plot option for your individual moraine exposure ages. Particularly for moraines with 2 or 3 exposure ages, these plots do not provide a good visual to see how your ages are distributed, nor the uncertainties on each age. Consider replacing the box and whisker plots with one-dimensional scatter plots (i.e., there is no vertical axis and ages are plotted as symbols with error bars). I recommended in the “Specific Comments” section that you should use mean exposure ages and standard errors for moraine age estimates, rather than the median age and IQR, so a visual of the IQR will be less relevant anyways.
Line 295: The statement about AAR method being commonly used in glacier reconstructions needs a reference or two to back it up.
Figure 4 caption: Consider adding a quick explainer for what “TLower”, “Smb”, and “AAR” stand for in your OGGM runs. I had to go back and forth to the figure as I was reading to figure out which was which.
Line 334: Should this cite be SM Table 2?
Line 360: Should this read “…a lack of moraines up valley of the latest Holocene moraines…”? To me, down valley suggests moving farther away from the glacier front, not back towards it.
Line 363: The European records feel somewhat out of place here. You are focusing on other glacial chronologies in northwestern North America, which all presumably were subject to similar climatic forcings, but glaciers in the Alps would have been subject to far different climatological influences. However, the similarities are certainly interesting! Consider moving this to later in the discussion, maybe a dedicated few sentences or paragraph about similarities between glacier chronologies in northwestern North America and elsewhere in the world that show the prevalence of the LIA throughout the northern hemisphere.
Lines 379-387: Much of this paragraph feels like important background information that should have been introduced earlier. This could be moved to a dedicated background section.
Line 382: Grammar edit – “…limitation to the AAR and THAR methods is that they do not account…”
Line 389: This paragraph feels like it is missing a topic sentence. I’d suggest a line summarizing the key points of your ELA reconstruction and what they imply about climatological changes since the LIA.
Lines 411-414: The first two sentences of this paragraph read like background material.
Supplementary figures 1, 4, and 7 and table 4 are not referenced in the text
Citation: https://doi.org/10.5194/tc-2023-55-RC1 -
RC2: 'Comment on tc-2023-55', Alia J. Lesnek, 26 May 2023
reply
In this manuscript, Hawkins et al. present a new 10Be chronology of late Holocene moraines in northwest Canada. They also reconstruct past regional climate from modeled glacier ELAs, simulate glacier volume over the past millennium, and forecast future glacier volume over the next century. The topic of this manuscript is well-suited for publication in The Cryosphere. In all, there’s an impressive amount of work that went into this study. The results and interpretations will be a useful contribution to our knowledge of past and future glacier change in Canada.
That said, the manuscript would benefit from revision before publication. Below, I've provided a few 'major' suggestions for the authors to consider; these are not all that major in the grand scheme of things, but may require substantial revisions to the text and some figures. I've also included a number of minor suggestions, indicated on a line-by-line basis. I hope that my comments offer constructive ways to improve the manuscript.
Major comments
Interpretation of 10Be ages: The pre-LIA exposure ages in your chronology are under-discussed in my opinion. I recognize that the number of exposure ages that you have per moraine is somewhat small and you are probably wary of over-interpreting your data, but if you assume that the late Holocene moraines date to the same event (which it seems you do based on your ELA reconstruction and modeling approaches), then I think you can expand your discussion of the exposure ages in at least two major ways.
First, the fact that you have erratics with consistent 11 ka exposure ages just outside of the late Holocene moraines suggests to me that the moraines were emplaced by a glacier readvance rather than a stillstand. Young et al. (2013), QSR have a nice discussion about using exposure age distributions to infer that a readvance created some of the Fjord Stade moraines in western Greenland. The distinction between a readvance and a stillstand to create the late Holocene moraines is an important one for your modeling and paleoclimate interpretations, and I think your dataset does allow you to distinguish between these two scenarios.
Second, given that these glaciers likely readvanced to deposit the late Holocene moraines, it seems reasonable that the boulders with exposure ages between 1-4 ka were reworked and therefore have small, but variable amounts of 10Be inheritance, and that the younger cluster of exposure ages more closely constrains the timing of moraine abandonment in your study area. That said, it’s also possible that the younger exposure ages are too young due to snow shielding or exhumation (e.g., although there isn’t a lot of debris on the glacier termini now, I could potentially see a scenario where the moraines were ice-cored early in their history and the youngest exposure ages are actually reflecting the timing of moraine stabilization in the region). However, to my eye, the tight clustering of young exposure ages across the six glacial systems combined with the scattered “old tail” of ages seems most easily explained by the inheritance interpretation.
These are just two areas where I think you can beef up your 10Be interpretations. A more thorough discussion of the above ideas (and other relevant ones that you may come up with during the revision process) as they apply to your study area would support the modeling you do later in the paper and strengthen your overall conclusions about glacier history in NW Canada.
ELA estimation methods: In line 407 of the paper and in the conclusion, you recommend that modeling modern glacier ELAs using climate data should be preferred over methods that rely on glacier geometry. I’m not fully versed on all of the latest glacier ELA literature, but this seems like an important outcome/recommendation from this study. Perhaps it’s outside the scope of this paper or it’s already been done by others, but I wonder how ELAs modeled based on climate data compare to ELAs calculated with in situ mass balance measurements. Are there any examples of glacial systems where this has been done before? Or even better, any glaciers in your study area with in situ mass balance measurements and corresponding ELAs that you could compare with your modeled ELAs to validate this approach?
Modeling details: The modeling exercises you did are a useful contribution, but the description of your model setups, particularly for the transient OGGM experiment, could use more detail. Perhaps I missed these things, but for example, where are you getting the starting values for glacier volume at 1000 CE? Are these coming from moraines or some other source? What ranges of Tbias and Pbias did you use to tune the climate models, and how do these bias values compare to what’s known about regional climate over the past 1 kyr? And more generally, why simulate the past millennium rather than some other time interval?
Figure 5: It would be helpful to include the known glacial history on this figure so the reader can see how the modeled glacier changes compare to the geologic constraints. I know this figure is showing ice volume for all 1235 glaciers in the study area, but a second panel showing something like a generalized time-distance diagram normalized to glacier length for the glaciers you studied in detail (incorporating data from the moraines and the satellite imagery) would help readers evaluate the results of your modeling.
Minor comments
Line 117: Can you include a supplemental figure showing your late Holocene glacier margins? Drawing paleoglacier margins can be quite challenging in the accumulation zone where there are no moraines, so this would be helpful to see.Line 145: Are the exposure ages presented in the main text calculated assuming no surface erosion or snow cover?
Line 178: Which glacier extents? Modern and LIA? What year did you use for your “modern” glacier extent (since you seem to have digitized glacier extents for many years from Landsat imagery)?
Line 235: This choice of mass balance gradient seems reasonable for a modern glacier, but I wonder how much of an impact that choice makes on your model output, especially since your glaciers advanced into the LIA?
Line 334: This SM table citation doesn’t look like it goes to the correct table.
Line 458: Reading this section and the last line of your conclusions, I’m left wanting a bit more information about the wider implications of glacier loss in this region. What specifically are the potential impacts on ecosystems? Are there people who rely on these glaciers for water, etc.?
Figure 2: This figure would be improved by adding your exposure ages and/or sample IDs so your readers can see how the ages are distributed across the moraines. Are the erratic boulders also included on this figure? It’s hard to tell.
Figure 3: Can you also show the individual exposure ages on the kernel density plot below the red summed curve? That will be useful for seeing how many ages contribute to a particular peak.
Figure 4: Define the abbreviations Tlower, Smb, etc. in your caption and in the relevant methods sections. I didn’t see these terms defined until the discussion.
Table 1: Exposure ages should be rounded to the nearest decade.
Citation: https://doi.org/10.5194/tc-2023-55-RC2
Adam Christopher Hawkins et al.
Adam Christopher Hawkins et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 176 | 85 | 11 | 272 | 25 | 2 | 4 |
- HTML: 176
- PDF: 85
- XML: 11
- Total: 272
- Supplement: 25
- BibTeX: 2
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1