In the context of future climate change, to quantitatively analyze the effects of glacial runoff on droughts is of great significance. The topic of this manuscript is important; however, this manuscript, in its present form, still contains several weaknesses. My major concerns are as follows: First, there is a lack of recent references (NONE published in 2021). This manuscript will be less persuasive without a comprehensive literature review. I would suggest the authors review and cite more relevant papers. Second, the authors use SPEI (rather than SPI) as an indicator of drought considering runoff. Although they have stated in Appendix C that “SRI has a moderate to strong positive correlation with SPEI”, the median correlation is only approximately 0.5, which is not high. I would suggest the authors further discuss the influences of using SPEI on the results. Otherwise, the conclusions are not reliable enough. Third, a big problem of this manuscript is that it does not take into account the CO2 emissions during the processes of melting glaciers. If just considering melting glaciers as the water source of runoff but ignoring the fact that it is also the emission source of greenhouse gas, the whole conclusion will be embellished. The authors should pay attention to this issue. Fourth, the authors are encouraged to further compare the results under the RCP 4.5 scenario against those under the RCP 8.5 scenario, although there are some relevant sentences in Lines 222-232. The potential findings may provide a reference for better policy making of drought mitigation. Some specific comments: 1. The authors are encouraged to summarize the innovation/major contributions of this study in the last paragraph of the Introduction part. 2. Eight GCMs are used in this study. Which eight GCMs? The authors do not mention them in the text. Please clarify this clearly. 3. Lines 86-87: “… we use a relatively short 3-month integration timescale, which is typical of that used to assess streamflow drought.” Any references for this statement? 4. As mentioned in Line 243, “current glacier meltwater production is unsustainably high”. I would suggest the authors discuss more about the possible results under such situation. 5. Appendix A: the authors only show the results for all the 56 basins in Figure A1, but do not provide in-depth discussions/explanations. For example, for some basins, blue shades and orange shades are overlapped; for others, blue shades and orange shades are not overlapped. Is it consistent with the result of K-means cluster analysis? Moreover, it is better to provide sufficient physical interpretations.
We thank the reviewer for their consideration of our manuscript. The reviewer raises interesting points, which we intend to address in a revised version of the manuscript at the close of the discussion period. At this time we would like to clarify two of the reviewer's list of four major concerns.
Literature review
The reviewer suggests that we ought to cite more relevant papers, especially references from 2021. At present, the manuscript includes 71 references, spanning the 1960s to the 2020s, with many published within the last 5 years. We believe this literature review is sufficiently comprehensive. If the reviewer has specific relevant 2021 publications in mind, we would of course be happy to include them.
Carbon emissions from retreating glaciers
The reviewer writes, "a big problem of this manuscript is that it does not take into account the CO2 emissions during the processes of melting glaciers. If just considering melting glaciers as the water source of runoff but ignoring the fact that it is also the emission source of greenhouse gas, the whole conclusion will be embellished."
The reviewer is correct that our manuscript does not account for a direct glacial effect on the global carbon cycle. Would the reviewer be able to direct us to a quantitative estimate of the greenhouse gas emissions potential of mountain glaciers?
Without a more specific reference, we are tempted to conclude that this is not a relevant consideration for our manuscript, on the following grounds:
A 2015 paper by Hood et al suggests that 9.68 Tg (0.00968 Pg) dissolved organic carbon will be “liberated from glacier storage” in mountain glaciers by 2050. A further 78 Tg total particulate organic carbon, with uncertain bioavailability and therefore a less well defined pathway to entering the atmosphere, could be released from all glacial systems. Note that this estimate of POC includes both ice sheets, which according to Hood et al account for 96% of the global POC store. So, a generous estimate, assuming that most glacial-derived DOC and POC come from mountain glaciers and are transformed directly into greenhouse gases, would be that glacial runoff is a net source of 87.7 Tg carbon to the atmosphere over the 35 years 2015-2050. Assuming constant flux over time yields an annual flux of 2.5 TgC. By comparison, the estimated annual net flux of carbon to the atmosphere from land use change during the 20th century was 2.0 Pg, or 2000 Tg (Houghton, 1999). That is, the estimated flux of carbon that could be liberated from glaciers is less than 1% the annual estimated flux of carbon from land use change alone during the last century.
A 2019 paper by Wadham et al, focusing on the role of ice sheets in the global carbon cycle, notes that DOC and POC fluxes from ice sheets are less than 1% the global land-ocean fluxes of the same. They further note that the microbial conversion of glacially derived DOC and POC to greenhouse gases is poorly constrained and subject to substantial uncertainties. A “conservative estimate” of methane flux from ice sheet runoff is stated as <1 Tg C per year. Once again, this flux is small compared with other global sources of carbon to the atmosphere.
The carbon budget produced by the World Meteorological Organization (https://public.wmo.int/en/resources/bulletin/annual-global-carbon-budget) estimates growth in atmospheric carbon of 6.2 GtC. Note that 1 Gt = 1 Pg = 1000 Tg. A generous assumption of 2.5 TgC directly from glaciers to the atmosphere (as above) would comprise 0.04% of the annual growth in atmospheric carbon concentration. Thus, it is unlikely that explicitly accounting for greenhouse gas emissions from glaciers themselves would have a first-order effect on 21st century climate and resulting glacial drought buffering.
We are aware of a 2021 paper by Fell et al., “Fungal decomposition of river organic matter accelerated by decreasing glacier cover”, the popular press coverage of which suggested that deglaciation could “speed up carbon emissions”. However, the article focuses on fungal carbon cycling and does not provide a quantitative estimate of greenhouse gas emissions potential from mountain glaciers—nor a comparison with other global sources of carbon. The conclusions of that article are therefore insufficient to suggest that glacially-derived greenhouse gas emissions will affect the results presented in our manuscript.
References in this comment:
Hood et al (2015). Storage and release of organic carbon from glaciers and ice sheets. Nature Geoscience 8: 91-96. https://www.nature.com/articles/ngeo2331
Houghton (1999). The annual net flux of carbon to the atmosphere from changes in land use 1850-1990. Tellus B: Chemical and Physical Meteorology, 51(2): 298-313. https://www.tandfonline.com/doi/pdf/10.3402/tellusb.v51i2.16288
Candela and Carlson (2017). The annual global carbon budget. Bulletin of the World Meteorological Organization 66:1. https://public.wmo.int/en/resources/bulletin/annual-global-carbon-budget
Wadham et al (2019). Ice sheets matter for the global carbon cycle. Nature Communications 10: 3567. https://www.nature.com/articles/s41467-019-11394-4
Fell et al (2021). Fungal decomposition of river organic matter accelerated by decreasing glacier cover. Nature Climate Change 11:349-353. https://www.nature.com/articles/s41558-021-01004-x
This study, by Ultee and co-authors, uses the results of existing GCM simulations and global glacier modelling to assess the importance of glacier runoff for drought buffering in the SPEI (Standardized Precipitation-Evapotranspiration Index). The SPEI is modified to quantify drought buffering by glacier runoff for 56 glacier catchments and compared with a baseline that does not include glacier runoff. The rationale for this work is outlined clearly within the introduction, and one of the key findings - that glacial drought buffering might extend beyond the end of the century despite passing glacier basin peak water for many regions - is a fascinating one that warrants further study. Overall, I found this to be a well-written and well-devised study that will be of benefit to researchers working on glacier-fed water security, and I identify no major issues with the manuscript.
Specific comments
Abstract - I think the abstract currently does a good job of explaining the “bare bones” of the study, but would benefit from some additional overview of the methods employed and key outcomes.
Line 110 – This is probably just something that I’ve seen and thus can never unsee, but you could consider changing SPEIw to SPEIG as currently it looks a bit like ‘spew’ and ‘G’ seems more representative.
Line 106 – could the title of section 2.1 be a little more snappy?
Line 144-145 – I’d suggest writing out the four variables here for clarity rather than having to refer to Table 1 which doesn’t appear for a few pages.
Lines 147, 169, and 188 – This could be down to personal preference, but while I appreciate the idea of using headline results as section headers, I wonder whether this would be better changed to a description of the wider content in each section. E.g. “Impact of glacier runoff for basin water supply”; “Glacier runoff influence on drought severity and frequency”, and “Influence of glacier cover for drought buffering”.
Figure 1 – I’d include the region / country name for each panel for ease of comparison. I also wondered whether it might be worth exploring changing one (or both) of the colours, as where they overlap it produces a brown-ish colour that’s not so dissimilar to that for the “without glacier runoff” scenario. Maybe changing the colour would still result in a bit of a “muddy” overlap, in which case stick with what you’ve got.
Figure 4 – the blue and green colours here are quite dull on my screen (the median gets particularly lost in the blue box plots), so might warrant brightening up for ease of viewing.
Figure A1 – as for Figure 1, I’d suggest including region names within each panel, and/or consider grouping the panels in this large figure by region, to allow easy intra-region comparison.
Lines 293-312 – because the SPEI is such a central component of the methodology here, I’d consider moving this detail into the main manuscript methods.