the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Apr 2024
Understanding microbial sourcing in Greenland subglacial runoff
Abstract. The microbial ecosystems that lie beneath ice sheets can impact and contribute to global biogeochemical cycles, yet remain poorly understood given the logistical challenges in directly accessing the subglacial environment. Studies instead often rely on indirect sampling of subglacial systems via the collection of meltwaters emerging from ice margins. However, the origin of exported material in these waters will change over a melt season as glacier hydrology responds to changes in surface melt. Here, we reveal trends in microbial sourcing (source environment) and assemblages in a large proglacial river in southwest Greenland by investigating three microbial datasets (16S rRNA) collected during different hydrological periods over three separate summer melt seasons. By combining microbial data with high-resolution hydrological and hydrochemical measurements, we show that changes in microbial assemblages follow changes in hydrological periods, likely influenced by variations in glacial drainage expansion inland with concomittant variations in inputs of surface melt and subglacial sediment exports. We further illustrate how relative changes in microbial assemblages can inform on the state of the glacial hydrological system, and also focus on methane-cycling populations to infer their potential distribution beneath the ice. Overall, our results highlight that timing matters when sampling proglacial rivers and we caution interpretations of exported assemblages without a good understanding of the catchment and system studied; this is especially true for larger systems which undergo more complex hydrological changes over a melt season.
- Preprint
(1796 KB) - Metadata XML
-
Supplement
(8501 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
CC1: 'Comment on egusphere-2024-817', Eric Boyd, 18 Jun 2024
Publisher’s note: this comment is a copy of RC1 and its content was therefore removed.
Citation: https://doi.org/10.5194/egusphere-2024-817-CC1 -
RC1: 'Comment on egusphere-2024-817', Eric Boyd, 18 Jun 2024
The paper describes chemical, hydrological, and microbiological data from coordinated samples collected from a proglacial environment draining Leverett Glacier, Greenland. The samples are broken out based on time of sampling and hydrological regime for the glacier, inferred from hydrographs and chemical data. This is then used to suggest drainage patterns and sourcing of fluids all of which is used to suggest that the time that a proglacial river is sampled will be reflected in different microbial community compositions, referred to assemblages by the authors.
Intuitively, this should be the case as oxidized meltwater enters the subglacial system and delivers oxidants to drive metabolism and also flushes out sediments that likely overwintered under low (no) flow conditions that promote anoxia to develop. The authors suggest that they can see this in their microbial data. Again, while this is perhaps the expectation, I have my reservations about several of the phylotypes/OTUs identified, most notably those identified in 2018 (e.g., Flavobacteriales) that were not identified in 2017 or 2015. Likewise, no Lysobacter in 2018. This would factor in largely to the patterns observed yet it is also possibly artifact due to the use of other primer pairs in 2018 than in prior years, the use of additional purification steps in 2015, and the use of different sequencing centers to generate these data. Other OTUs (e.g., Rhodoferax, Polaromonas) perhaps behave more predictably (slight increase then decrease during hydro cycle) and are more cosmopolitan across years. I suggest that the authors remove OTUs that are not present across years and reperform their analyses to see if they still see the patterns emerge at a compositional level. Otherwise, strong statements about cause and effect need to be better tempered, in particular in the abstract (lines 28-31) and conclusion (line 523). Further, based on the methods and locations of sampling, I don't think that a definitive statement that the same site was sampled throughout the melt season can be claimed (line 522).
Other more specific comments
Lines 16-18: How can one make this statement if these systems are so poorly understood. My recommendation - delete the first part of the sentence or reword to indicate that their contribution to biogeochemistry is poorly understood due to logistical challenges.
Lines 19-20: I don't disagree but as far as I can tell, this is the hypothesis to drive the work rather than it being definitively known.
Line 77 and throughout: I imagine geochemists/hydrologists will read this and wonder what is meant by assemblages in this context. Please refer to assemblages more descriptively (e.g., microbial assemblages or communities)
Lines 158-159: Given the effect of minerals on DNA recovery and purity, and downstream effects on PCR, i cannot imagine that this additional purification did not effect the outcomes. Can the authors demonstrate that this has an effect (or not) using an existing sample, one of which is purified like this versus not and then sequenced to determine the introduced bias. I also wonder why replicate subsamples were not sequenced but were rather pooled, at least for one sampling point to assess variation that may exist?
Lines 190-191: These would seem to be possibly informative since they too could point to input of surface materials and photosynthetic biomass/organisms.
Line 209: I don't understand this. Perhaps inserted due to a prior reviewer comment? Regardless, the authors did about as much with their putative OTUs involved in CH4 cycling than they could have.
Lines 216-218: 'Candidatus Methanoperdens nitroreducens' and then 'Cand. M. nitroreducens' thereafter.
Figure 2c and 2d. Does this then suggest that the drivers of community composition are different depending on the time by which you sample them? In other words, is this stochastic (mixing of different source waters) rather than deterministic (true environmental drivers) that are difficult to deconvolute? This seems to be likely as i cannot imagine that what the authors are seeing is community turnover but rather is the former (mixing of different communities that themselves are deterministically structured).
Figure 2 and more generally: Can the authors estimate biomass loads from their DNA recoveries and the volumes of material recovered. It would be interesting to see how this might have changed over a seasonal melt cycle.
Lines 319-321: But the read number is irrelevant correct, since the sequence library was downsampled prior to calculation of alpha diversity, correct?
Lines 400-401: This could be tempered based on above comments.
Lines 406-407: I would imagine these would also be the most anoxic of the sample regimes and the community composition would reflect this. Yet, the communities are dominated by putative aerobic methanotrophs...
Lines 419-421: A good place to include discussion of mixing and stochasticity and determinism and their relative importance in explaining what is observed.
Line 460: Again, I would be looking for chloroplast sequences. Were Cyanobacterial sequences identified at all that could help?
Lines 492: I don't understand this - the reduced O2 availability would likely select against methanotrophs and rather favor methanogens. Are the authors suggesting that the low O2 is due to consumption? if so, then why not complete consumption of methane
Lines 508: Anaerobic methanotrophs and aerobic methanotrophs don't share a niche. Their niche might partially overlap in that they both oxidize CH4 but O2 will partition them beyond that. Suggest modifying this statement to avoid confusion.
Citation: https://doi.org/10.5194/egusphere-2024-817-RC1 -
RC2: 'Comment on egusphere-2024-817', Anonymous Referee #2, 28 Jun 2024
As a microbial analysis of several years of Leverett Glacier samples, “Understanding microbial sourcing in Greenland subglacial runoff” provides an excellent look into the variability of microbial communities in emergent Greenland Ice Sheet waters. The combined data from multiple years and multiple hydrological regimes shows a far more complex picture than a single shorter study could. I have several suggestions for improving the manuscript that should be possible to address in the course of revisions. In general, I think the introduction needs work and several points in the discussion are not necessarily well supported. The supplemental tables also need work.
Line 38: Consider a word other than “leak”, which implies built infrastructure. “Evade” perhaps?
Lines 41-50: I don’t think marginal sampling should be presented as some sort of work around. The glaciohydrological system naturally samples the subglacial environment, and I don’t think anybody has ever seriously suggested that what emerges from a glacier differs importantly from what’s beneath it. The very first work that demonstrated widespread bacteria in the glacial bed (Sharp et al., 1999) combined samples collected in boreholes with marginal samples, and did not find an appreciable difference in the bacterial content. There are a number of studies that have directly accessed the subglacial environment in Antarctica (oddly are not cited here). But in Antarctica, marginal sampling of subglacial waters is generally not possible, as the vast majority of subglacial waters discharge into the ocean under ice shelves. Thus, Antarctic biogeochemistry is largely limited to drilling projects; whereas in Greenland and alpine settings, nature can do the sampling for us. Not to say drilling projects aren’t useful or interesting in these settings. Graly et al. (2014) find some differences in water chemistry between boreholes and the integrated outwash at Isunnguata Sermia, and similar things have been observed in alpine glaciers (e.g. Tranter et al.). But all of these differences have glaciohydrologic explanations (i.e. boreholes being more isolated, etc.). I think this section should be reframed and probably expanded to a broader selection of literature.
Lines 52-65: This paragraph should probably cite more literature. There were quite a few insights into glaciohydrology developed between the theoretical work of the early 1970s and Davidson et al. (2019). There’s a certain facile element to the framing/presentation here. The notion that temperate glaciers change seasonally is not exactly novel. Maybe the references cited in lines 87-89 could be moved forward to this paragraph, but with an actual explication of their findings.
Line 64: Be more precise than simply saying “older”. In a geologic sense, the bed material is billions of years old at this site. This is discussing the sediment residence time, i.e. the time between when subglacial materials enter the glaciohydrologic system and when they leave it.
Line 68: I would avoid using the abbreviation LG. Non-standard abbreviations have the potential to confuse readers. It is not needed for space reasons. I would also avoid GrIS. OTU might be okay, simply because it is used so frequently. But LDA, LEfSE, PCoA, nMDS, etc. become a garble to those not deeply familiar with these methods.
Line 76: I would use glaciofluvial rather than fluvial here. The uninformed reader might think you are referring to a normal subaerial river.
Lines 80-84: The methane question should be more properly developed, if this is a major aim of the paper. I.e. disentangle this list of references to state what’s known and unknown about methane cycling in this sector of the Greenland Ice Sheet.
Lines 87-89: As I mentioned above, these references should be properly presented in the introduction.
Line 92: I might say “a main source” rather than “the main source”. The branch of Akuliarusiarsuup Kuua that drains Russell Glacier and the lateral margin of Isunnguata Sermia is also pretty significant.
Line 150: Do you have any sense of the quantity of glacial surface runoff that makes it directly into the proglacial stream at Leverett Glacier? This could, in principle, be significant and seasonally variable.
Lines 237-238: This seems more a discussion type statement than a results heading.
Lines 237-296: I wonder if this should be results at all; it’s mostly an elaborate description of the site context. Maybe move forward and give it its own heading. Or incorporate into an extended section 2.2 of the methods.
Line 263: I might consider a different term than “outburst period”. To me, “outburst” implies something like a lake drainage event, whereas this is the normal spring escalation of the glaciohydrologic system.
Line 319: A section header is very much missing here! This is no longer about “hydrological evolution”.
Figure 3: Is there a good reason for only plotting the top 5 here? I would be interested in tracing all of them, assuming the data is usable / significant. I also don’t believe the data underlying this figure is in the supplemental tables (unless I some how missed it). Supplemental table 5 has a list of relative abundance of each of the OTUs, but I think in total for the whole suite, not by sample.
Supplementary Tables: I found it difficult to understand the supplemental tables. The names of the sheets don’t match the labels on the tabs. Some of the data is not legible without expanding the columns, and the meaning of some of the headings is not entirely clear.
Line 397-400: What about non-subglacially routed surface water? Did you do anything to look at the amount of water coming off the surface or off the ice-cored lateral moraines? This might be especially relevant before subglacially routed flow picks up.
Line 409: Again, I don’t particularly like the framing of these as “outburst events”. You could just call them “spring events”, like you do on the 2018 panel of figure 1.
Line 414: Is lysobacter really a “putative anaerobic” taxon?
Lines 419-421: I’m not entirely convinced of this statement. What makes this period unique is the abundance of OTU 3, 4, and 6 (at least on Fig 3). To me these seem more like soil bacteria than a deep anaerobic source. Again, surface runoff could have been a more significant component in the early season. Whatever the source, it’s a supply that is entirely or nearly entirely absent in different conditions.
Lines 436-440: This list doesn’t match with the late melt highlighted taxa in fig 4. The OTUs should be 2,5,8,13, and 25 (i.e. not 1 and 4, which do not appear to increase here, but 8 and 13 instead). And again, I’m not entirely convinced of the diagnosis of this as a fingerprint for the ice sheet surface.
Line 452: Typo: “appart”
Lines 478-483: This might cast some shade on the notion that the “2017 late-melt” is a surface signal. Or maybe OTU 7 has some very particular circumstances under which it flourishes.
Line 492: Is there a figure limit in this journal? S3 seems quite important, so I would make it part of the manuscript proper if allowed.
Citation: https://doi.org/10.5194/egusphere-2024-817-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 218 | 79 | 18 | 315 | 221 | 9 | 9 |
- HTML: 218
- PDF: 79
- XML: 18
- Total: 315
- Supplement: 221
- BibTeX: 9
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1