| 22 Jan 2026
Sensitivity of Andean Glaciers to ice-flow parameters in the Parallel Ice Sheet Model
Abstract. Mountain glaciers are losing mass rapidly due to anthropogenic climate change. Projections of glacier evolution across the Andes under different warming scenarios have primarily been as part of global scale modelling frameworks, rather than dedicated, regionally optimised, simulations. These global-scale models use simplifications of ice flow physics that may be unsuitable for steep topography, such as that which occurs at mountain valley glaciers. More complex models are available, but with that complexity comes further sources of uncertainty. Here, we assess the sensitivity of the Parallel Ice Sheet Model to ice-flow parameters influencing the ice rheology and subglacial sliding characteristics. We find that the resistance of subglacial material has the most impact on modelled ice outputs (e.g., ice volume), followed by the exponent which relates basal shear stress to sliding, and the threshold velocity at which sliding occurs. The ice-flow rheology enhancement factors, the rate of subglacial water decay, and the maximum water thickness within a presumed subglacial drainage network, can either cause minor variations, or no effect at all, on ice outputs. Our study informs what parameters can potentially be negated in future parameter ensemble tests and provides direction on where further investigation is needed.
General comments
The present manuscript aims to test the impact of various ice-flow parameters of the Parallel Ice Sheet Model (PISM) over numerically modelled glacier outputs in the Andes. This study is motivated by a research gap on the appropriate parameterization of glaciological parameters in PISM over mountain glacier settings, the simplistic nature of global-scale models in replicating accurate ice flow, and the importance of the study area for water resources under future climate change projections. The manuscript addresses the aim through two modelling experiment ensembles, iterating enhancement factors, subglacial properties- and basal-sliding-related PISM parameters through default, minimal and maximal values over five Andean hydrological catchments, assessing their effect on ice volume and area. The study finds that some parameters, such as enhancement factors (E), the rate of subglacial water decay (C), and maximum water thickness (Tm) have little-to-no impact on the modelled outputs, whereas some of the subglacial properties-related ones, namely the resistance of subglacial material (ϕ), the exponent relating basal shear stress to sliding (q), and the sliding velocity threshold (𝑈𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑) have a significant or higher impact. The authors conclude that such findings can help in the selection of the appropriate ice-flow parameters in future work, improving both research design and conservation of computational resources.
Overall, this manuscript represents a significant contribution to our understanding of how several PISM ice flow parameters influence ice volume and area outputs, which have not been intensively experimented with in mountain glaciation settings before. Thus, I foresee this work being a future reference point for many PISM users, who often resort to default parameter values or guesswork in their choices when modelling the (palaeo)climate conditions favourable for empirically constrained glacier extents. Without accounting for uncertainty/sensitivity of model parameters, results are less robust and convincing, which this work emphasizes through the variability exhibited in several ice volume outputs. Additionally, as many users encounter inefficient use or estimation of required computing resources in High Performance Computing access applications, this work will help streamline the workflow and prioritise the most relevant sensitivity tests. Furthermore, I commend the authors for packaging this work into a concise, well-structured and -illustrated manuscript that is mostly easy to follow. The Zenodo repository files further provide useful PISM scripts that can be checked by interested readers and model users. The Supplementary Information includes clear tables and mapped model outputs that support the manuscript interpretations and conclusions.
Specific comments
My main comment pertains to embedding more literature in the explanations of the chosen parameters and their behaviour. While this is done well at times, the manuscript could engage with more empirical and/or modelling studies that would support their choices and arguments, or when it does not exist, to explain how such choices were made (see in-line comments). Furthermore, I would encourage the authors to emphasize the potential impact of size and climatic zones of the model domains on the ice-flow parameter sensitivity tests – while the former is acknowledged throughout the results and interpretations, there is no unifying conclusion emerging out of it. In the case of the latter, the climatic zones are described earlier in the manuscript (lines 72-79) but their potential implications (or lack thereof?) are not addressed. Therefore, I think a small subsection/paragraph(s) on these aspects could enrich the manuscript.
Technical corrections
Below are some further in-line and Figure/Table comments that include minor suggestions, changes and considerations for the authors to include in the manuscript or consider in their upcoming or future work.
Lines 54-55 – perhaps it is better to be specific about the exact kind of statistical analyses undertaken (these are revealed later, but might be good to explicitly state them here)
Lines 65-67 – this long phrase could benefit from clarity – perhaps write it as two sentences
Figure 1 – clear, nicely labelled maps. One minor suggestion about the DEM colour scheme– at first sight, it is difficult to tell which is higher/lower ground (in the case of the zoomed-in maps). In my opinion, this could be improved by underlaying them with a hillshade model or using a more traditional topographic colour scheme for the DEM (brown-high – green-low). Otherwise, including the elevation label for each zoomed-in map might also work well.
Table 1 – Table is organized very clearly. It might also be useful to include a column with references for each parameter. I am also wondering where the minimum and maximum value for the parameters have been determined from – this could go either in-text or in the table.
Lines 123-124 – it might be useful to explain the variation between the min and max value, both in terms of values and intervals, justifying the rationale
Line 126 – is it called T component because of Till? For the sliding (S) and enhancement (E) parameters, this seems logical – perhaps a mention either in the heading or text would help.
Lines 128-133 – this paragraph is explained very well, but I feel some references might be needed here, e.g., on what basis do you know about the thick sediment in the glacier forefields? Are the glaciological explanations supported by other studies?
Lines 184-185 – could perhaps explain how the choice of the number of ice and bedrock layers was determined – was it by trial or error or suggested by previous work? Are there any effects of this choice on the model outputs and what would these potentially look like?
Line 186 – what is not clear at this point for the reader is whether the mountain glaciers were grown from no ice conditions – perhaps this can be explicitly stated.
Lines 200-205 – this paragraph seems vague – what do you mean by ‘commonly used and extended ranges’? Although the ensemble design becomes a little clearer as the reader progresses through the manuscript, a table exemplifying the combination could be useful – there are some examples in the Supplementary Information that could be cited or brought forward as a table here.
Lines 224-225 - ‘We used Davies (2013) which uses…’ - try to avoid repetition
Lines 230-232 - very good acknowledgement of the WorldClim dataset limitation regarding underestimation of temperature in mountain peaks. Perhaps in future work, the authors could consider the CHELSA dataset (Karger et al., 2017; 2023). It is a downscaling of ERA-interim climatic reanalysis. Although proven to resolve temperature similar to other climate datasets, it yielded more accurate results for precipitation (Karger et al., 2017).
Lines 244-245 – Although somewhat understandable, this sentence could benefit from some clarity/simplification
Line 251- ‘varying’ and ‘varied’ – consider avoiding repetition in the same sentence
Lines 335-337 – this phrase is slightly long and convoluted – it might need simplifying.
Line 362 – remove second ‘.’ After ‘difficult’
Lines 363-365 – good explanation as to why some of the parameters do not have a sufficient effect on the modelled ice outputs. Regarding the PISM hydrology, it might be good to briefly add something here about the default hydrological model and its alternatives (I am aware experimenting with different hydrological models is a suggestion for further work later in the manuscript)
Lines 378-380 – excellent explanation the different ϕ values but might be useful to include some references here
Line 392 – ‘simulations’ instead of typo ‘simulation s’
Line 393 – do you mean ‘centered’ instead of ‘cantered’?
Lines 393 – 395 – good interpretation of the strong influence of ϕ on ice volume outputs – I am wondering whether highlighting this in envelopes on Figure 7 could make them more obvious to the reader? Furthermore, you use catchment numbers in the text (e.g., #4, #5), but these are only labelled with letters on Figure 7 (A-E). Consider adjusting either on the Figure on in the text.
Lines 397-399 – keep font consistent in the final submission
Lines 481-482 – typo ‘parameters mention’ instead of ‘mentioned’. However, it might be better to recap the names of the specific parameters instead
Lines 497-500- good to see appropriate recommendations for sensitivity tests. I would just add that not only is the PISM PDD model sensitive to climatic parameters, but so is the diurnal energy balance model simple (dEBM-simple). This model was argued to be an improved alternative to the PDD model, by accounting for the melt-albedo feedback, without significantly increasing computational time (Zeitz et al., 2021; Garbe et al., 2023). However, to the best of my knowledge, this model has only been applied in ice sheet settings. Therefore, it could represent a research gap for mountain glaciation applications, and it might be worth mentioning as a recommendation for future work.
Line 535 – zenodo link is not easily clickable/searchable – please adjust the link (this one worked: https://zenodo.org/records/17878115 )
Suggested references:
Garbe, J., Zeitz, M., Krebs-Kanzow, U., & Winkelmann, R. (2023). The evolution of future Antarctic surface melt using PISM-dEBM-simple. The Cryosphere, 17(11), 4571-4599. https://doi.org/10.5194/tc-17-4571-2023
Karger, D. N., Lange, S., Hari, C., Reyer, C. P., Conrad, O., Zimmermann, N. E., & Frieler, K. (2023). CHELSA-W5E5: daily 1 km meteorological forcing data for climate impact studies. Earth System Science Data, 15(6), 2445-2464. https://doi.org/10.5194/essd-15-2445-2023
Karger, D.N., Conrad, O., Böhner, J., Kawohl, T., Kreft, H., Soria-Auza, R.W., Zimmermann, N.E., Linder, H.P. and Kessler, M., 2017. Climatologies at high resolution for the earth’s land surface areas. Scientific data, 4(1), pp.1-20. https://doi.org/10.1038/sdata.2017.122
Zeitz, M., Reese, R., Beckmann, J., Krebs-Kanzow, U., & Winkelmann, R. (2021). Impact of the melt–albedo feedback on the future evolution of the Greenland Ice Sheet with PISM-dEBM-simple. The Cryosphere, 15(12), 5739-5764.https://doi.org/10.5194/tc-15-5739-2021