the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Inter-model differences in 21st Century Glacier Runoff for the World’s Major River Basins
Abstract. Projected glacier change has important downstream consequences, including sea level rise, changing freshwater supply, and loss of important cultural sites. While the glacier contribution to global sea level rise and associated uncertainties have been quantified in model intercomparison studies, comparatively less focus has been directed towards the inter-annual changes in runoff caused by glacier recession. The observed effect of glacier runoff on basin-level water availability makes simulated future runoff a particularly consequential target for analysis. In this study, we compare century-scale runoff simulated by three global glacier evolution models. Aggregating annual glacier runoff contributions to 75 globally-distributed major river basins, we find that the three models agree closely in some basins but differ dramatically (up to a factor of 3.8) in others. However, when we analyze century-scale runoff changes relative to a glacier model's historical runoff baseline, annual runoff projections are much more consistent across glacier models. Glacier models project broadly consistent relative changes in seasonal runoff supply, with some differences across climatic regions. Estimates of the year of peak water are more consistent across glacier models (when driven by a climate model ensemble) than across individual climate forcings within a single glacier model. We identify the glacier models' different approaches to modifying precipitation forcing as the dominant source of inter-model differences in projected runoff. Our findings highlight the comparative roles of glacier evolution model, global climate model forcing, and emissions scenario as important sources of uncertainty across different metrics of projected glacier runoff. We recommend steps to account for glacier model uncertainties in glacio-hydrological modelling efforts.
- Preprint
(11033 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1778', Mauri Pelto, 29 Sep 2024
I appreciate the authors providing an article that triggered my curiosity and motivates a desire to know more, this is indicative of a useful contribution. The evaluation of three glacier models and a suite of GCMs provides a useful quantification of the variation and congruence of the resulting glacier runoff and peak water timing.
The comments below are directed at encouraging the authors to provide further context that will clarify the results and increase the impact. Responding to the comments are an opportunity not a requirement for publication of this study. Defining the seasonal/annual time interval of peak water identification is crucial. Further exploration of the reasons behind model variations, including referencing other related studies for basin comparisons if practical.
Specific Comments
18: For consistency in the referencing in this section and to focus on aspects that are not human centric, it would be appropriate to add an older reference for altering ecosystems to Bossons et al (2023), maybe Jacobsen et al (2012) or Pittman et al (2020).
26: Peak water can be defined seasonally or annually, in terms of the glacier runoff component or overall discharge. Is peak water evaluated on an annual or seasonal basis and is it for glacier runoff or encompassing of the area of the watershed initially occupied by the glacier?
152: Explain the rationale for using a 20-year rolling mean to determine peak runoff timing? This is a long-time interval.
182: An interesting distribution of highest relative runoff for regions for OGGM and GloGem, as there is not a shared climate or glacier type. Huss et al (2017: Fig. 3) and other have provided nice global maps of contributions of snow and ice melt vs effective precipitation revealing a wide range that illustrates a different picture based on specific hydrologic variables.
184: Moore et al (2020) utilized the glacier wastage contribution to runoff to distinguish glacier mass loss vs runoff from seasonal snow melt loss or precipitation. In looking at your results does this consideration explain any of the SSP variation? They observe that glacier-melt contributions have already passed peak water in the Columbia River headwaters. and there is a declining trend in realized streamflow.
185: It would be useful to provide a bit of a comparison for these five basins shown in Figure 2 and 3. % of basin with glaciers, mean annual precipitation range for glaciers, % of discharge during peak season provided by glaciers. Does not have to be these specific variables, but some that differentiate the basins.
192: The Yukon basin much different in terms of glacier size/type then Rhone or Glomaa. Another interesting geographic range in shared response that could be explored further.
Figure 4. This is an exceptionally valuable figure. The time frame/range of overlap is relatively narrow in most cases in Column b, not sure if it is worth trying to better illustrate this with a shading of some sort.
212: Why is the GCM peak water timing range so large compared to SSP and GM?
222: Does seasonality increase as you go further south in the Andes. If so that is worth noting.
223: Is the range of increase from arid to maritime basins indicative of the precipitation trend?
224: The increasing magnitude of peak runoff reported for the Skagit basin is not in agreement with other modelling or observations of glacier runoff. A peak water year of 2026 is noted, although peak flow has already occurred in this area according to (Riedel and Larabee, 2016; Pelto et al 2022) based on streamflow and glacier runoff observations in the region and modelling from (Bliss et al 2014; Moore et al 2020; Ultee et al 2023). Are the glaciers models utilized generating too much runoff from this region, or is this discrepancy the result of accounting for runoff from the same area, including former glaciated and currently glaciated areas as time progresses, and hence the models reflect increased precipitation/snowfall? If so it is worth providing a description of this using a specific use case. For maritime Alaska, the rising maximum peak runoff observation is supported by results from Young et al (2021).
Figure 6: This figure illustrates the variation by model, would be useful to add to this figure a geographic domain using same color scheme but adding symbol variation.
310: Reword –“are much more similar” to “significantly reduced offsets”.
320: Utilize a reference that has attempted to quantify this precipitation variability in GCM projections.
339: In terms of human intervention is probably worth noting key mechanisms, most importantly diversions for agriculture or hydropower and reservoir storage.
352: Given the quantitatively consistent earlier shift in runoff timing and lower magnitude of runoff, reported across glacier models. Comment further on the impact on drought buffering.
Bliss, A., Hock, R. and Radić, V.: Global response of glacier runoff to twenty-first century climate change, J. Geophys. Res. Earth Surf., 119, 717–730, doi:10.1002/2013JF002931, 2014.
Jacobsen, D., Milner, A., Brown, L., and Dangles O.: Biodiversity under threat in glacier-fed river systems, Nat. Clim. Change, 2,361–364. https://doi.org/10.1038/nclimate1435, 2012.
Moore, R.D., Pelto, B., Menounos, B. and Hutchinson, D.: Detecting the Effects of Sustained Glacier Wastage on Streamflow in Variably Glacierized Catchments. Front. Earth Sci., 8:136. doi: 10.3389/feart.2020.00136, 2020.
Pelto, M.S., Dryak, M., Pelto, J., Matthews, T., and Perry, L.B.: Contribution of Glacier Runoff during Heat Waves in the Nooksack River Basin USA. Water, 14, 1145. https://doi.org/10.3390/w14071145, 2022.
Pitman, K., Moore, J., Sloat, M., et al,: Glacier Retreat and Pacific Salmon, BioScience, 70 (3), 220–236, https://doi.org/10.1093/biosci/biaa015, 2020.
Riedel, J. and Larrabee, M.: Impact of Recent Glacial Recession on Summer Streamflow in the Skagit River. Northwest Science, 90(1), 5-22. https://doi.org/10.3955/046.090.0103, 2016.
Ultee, L., Coats, S., and Mackay, J.: Glacial runoff buffers droughts through the 21st century. Earth System Dynamics, 13, 935–959, https://doi.org/10.5194/esd-13-935-2022, 2022.
Young, J. C., Pettit, E., Arendt, A., Hood, E., Liston, G. E., and Beamer, J.: A changing hydrological regime: Trends in magnitude and timing of glacier ice melt and glacier runoff in a high latitude coastal watershed. Water Resources Research, 57, e2020WR027404. https://doi.org/10.1029/2020WR027404, 2021.
Citation: https://doi.org/10.5194/egusphere-2024-1778-RC1 -
RC2: 'Comment on egusphere-2024-1778', Pascal Egli, 16 Nov 2024
This study compares the performance of three glacier models fed with GCM and reanalysis data to assess the future impact of glaciers on discharge variability in 75 major river basins globally. It assesses the timing of peak-water, future trends in seasonality and glacier contribution for different climate scenarios, climate model configurations and glacier models.
The authors followed a rigorous methodology, studied uncertainty in a convincing way and presented interesting results, unveiling large discrepancies in the results for peak water between different climate model configurations and climate scenarios, but also significant differences between different glacier models for certain catchments.
The authors put in good efforts to compare the intrinsic differences between different glacier models by comparing precipitation inputs for the 2000-2019 period, by normalizing model results with the reference period and comparing glacier models for 75 different catchments.
Ultimately, the article highlights the need for more mountain hydrological observations to better set-up and constrain glacier evolution models and glacio-hydrological models.
I recommend the article for publication after minor adjustments. Addressing the following specific comments might be of help for improving the manuscript.
Specific comments:
L13: Sounds a bit vague. Which component contributes how much in uncertainty?
L15 runoff. “We recommend steps to account for glacier model uncertainties in glacio-hydrological modelling efforts.»
This could be formulated more explicitly.
Figure 1: “…basin’s average annual glacier runoff over the entire period” Why not show the percentage of runoff that comes from glacier melt? This would show how important glaciers are for each catchment. Absolute numbers may be less important.
The colors in this figure make it hard to distinguish between different catchments and their glacier runoff. A different, non-linear, color ramp could be appropriate (more distinct scaling for smaller numbers).L147: How were the 12 forcing GCM realizations chosen among the 250 realizations in the CMIP6 ?
L152: For calculating peak water, 20 years is a long time for a rolling mean, especially in the context where you consider the time scale of one century for the entire simulations. Could you elaborate on the choice of the 20-years’ time interval for the rolling mean?
L155: GCM range of peak water: How did you decide on calculating the range like this? Could it have been calculated in a different way?
182: What do these regions have in common to result in such similar projections for one model?
L185:
Figure 2: Very interesting figure. How did you select exactly these basins?
Maybe elaborate a bit more (in the text) why the models are so different for the Aral sea, and two of the models are very different for Serrano.198: A reduction in variability due to normalization is something that is to be expected…
Figure 4: This is a very instructive figure. Some of the GCM ranges of peak water are very large, and this is concerning when thinking about future scenarios.
211: Explain better why you do the rolling mean analysis, and why you chose 20 years as a time span.
212: This means that determining peak water timing depends largely on the climate model that is chosen. Meaning that there is substantial uncertainty in those results - not due to the differences in glacier evolution models, but due to the GCMs. (just in the case of basins with little to no evolution of runoff over time…)
Figure 5: Could these flows be put into relation with total (not just glacier) basin runoff? Both as a percentage and as absolute numbers... to illustrate how important glaciers are in each catchment – in the present, and in the future.
229: It would be helpful to your argument to show seasonal runoff also for the period 2040-2060.
248: "Glacier models have generally been calibrated to match per-glacier mass change observations...": This could be better explained.
257: Temperature variability and its impact on glacier melt could be discussed more, also related to large GCM variability. Significant differences in temperature for different GCM realizations or for different climate scenarios will naturally also lead to differences in partitioning between liquid and solid precipitation.
267: Here, it would be interesting to explain more about the influence of different melt parameterizations.
292: Which properties of GloGEM could be responsible for maintaining broader seasonal peaks, and which properties of OGGM lead to maintaining strong seasonal peaks?
335: The role of glacier melt could be discussed more in detail here and throughout the paper.
Figure A1: I cannot see the line indicating the year of peak water.
Citation: https://doi.org/10.5194/egusphere-2024-1778-RC2
Model code and software
Runoff Processing Notebooks Finn Wimberly and Lizz Ultee https://github.com/finnwimberly/glacial_runoff_intercomparison
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
427 | 107 | 87 | 621 | 21 | 25 |
- HTML: 427
- PDF: 107
- XML: 87
- Total: 621
- BibTeX: 21
- EndNote: 25
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1