General comments:

I very much enjoyed reading the preprint ‘Monitoring Arctic Permafrost – Examining the Contribution of Volunteered Geographic Information to Mapping Ice-Wedge Polygons’ by Walz et al. (2025). In the study the authors evaluate the potential of Volunteered Geographic Information (VGI) for mapping and monitoring permafrost by conducting a case study of ice-wedge polygons in two study regions, located in Alaska and Canada.

In general, the manuscript is well prepared and in well the scope of The Cryosphere. The purpose of the manuscript is clearly articulated; the research process is in most parts described sufficiently and appropriate methods were utilized. Although the approach of this study is not totally novel, the manuscript has added value to the use of VGI in permafrost monitoring. However, the discussion remains a bit superficial at some points and should be revised, for example, the broader relevance of the findings should be discussed more thoroughly. In addition, the structure of the manuscript could benefit from small changes. I’ll point out these shortcomings and gaps in the discussion in the next section ‘Specific comments’ – so the authors can take actions on them more easily.

Overall, the manuscript is clear and easy to follow, written in good English, and provides interesting insights into the use of VGI in mapping and monitoring of ice wedge polygons.

Specific comments:

• Line 13: Referring permafrost as a key climate variable should be considered again. I do not consider permafrost as a climate variable as its formation depends on the climate (and soil characteristics, moisture etc.) rather than it being a climate variable itself. Define permafrost properly.
• Line 125: Is it possible to provide more information on how much smaller the area is if the volunteers are asked to trace the polygon outlines compared to just point digitization of the polygon centroids? This could help future researchers to plan better the best use of VGI for their own studies.
• Clarify the use of different terms throughout the manuscript or use the same term consistently. In line 141 and in Fig. 1 in the first boxes from left, authors use the term “permafrost centroid”, whereas the terms “ice-wedge centroids” or “ice-wedge polygon centroids” are used. Be consistent to avoid confusion.
• Post-hoc test (Tukey HSD) was not mentioned in the methods section, but first time mentioned in Appendix C, when presenting the results of the test. Please mention this also in the methods section.
• Lines 167-169: Please clarify. Now it is unclear for the reader what does the “()rettelbach2021quantitative” refer to.
• Lines 176-178: Clarify: In which one of these mapping events were the majority of the points digitized? Or were there three events organized simultaneously in the three different schools and one university, and majority of the points were digitized during one of these events organized at the same time at multiple locations?
• Line 197: Provide the number of the centroids also for BH as they were provided for the CB couple of lines earlier.
• Lines: 207-208: Double check the distances, such figures are not presented in Table C1. This causes confusion, so please recheck, clarify, and/or revise Table C1.
• Fig 6. and its caption: Add legend for the colors in the table or remove different colors if they are not necessary. Now the reader does not understand why the cells are in different colors and it is confusing.
• Major comment #1: The structure of the manuscript could be enhanced as some of the text is currently misplaced. For example, the section “4 Results” contain also discussion of the results (e.g. lines 201-202, 216-219, 229-237, and 280-284). Authors should either combine sections 4 and 5 into one section “4 Results and discussion” or then the current Discussion section should be expanded to contain all discussion currently found along the results. In its current form the Discussion section is quite superficial and short. In addition, the results section contains also some description of the methods (e.g. lines 239-248). There is also some overlap between results and methods (compare e.g. lines 162-163 and lines 260-263). Provide all needed info in the methods section.
• Major comment #2: How does the experience of the volunteers affect the results? In the study volunteers ranged from 12-18 years old to university students of geography. It can be assumed that geography students would have much more experience and knowledge on permafrost, the ice-wedge polygons, and GIS data compared to younger students. Were the tailored teaching materials presented before the mapping events enough to balance the different skill levels of the volunteers? It would be interesting to see separately the accuracy of the different volunteer groups compared to the experts. This info could help researchers plan such mapping events in the future to get as accurate data as possible. This would also give more content to the discussion of the limitations and potentials of the VGI in permafrost monitoring. Authors could at least discuss the potential influence of having volunteers with different experiences to the accuracy of the data.
• Another possible topic to discuss is the different ways of acquiring remote sensing data. In CB aircraft data with resolution of 20 cm was utilized, whereas in BH drone data with ~10 cm resolution was collected. Does different RS data affect the mapping results? Which alternative is preferrable in future studies if one can choose?
• In addition, the applicability of the method for monitoring other permafrost landforms could be discussed shortly to provide broader perspectives of the potential/limitations of the VGI in permafrost monitoring.
• Line 325-326: Was this shown in this study or elsewhere? Please add references if needed
• Lines 329-331: Previously it was stated (in lines 216-219) that this result is in line with previous research (Herfort, 2018). Please add references also here. In general, there are relatively few references in the Discussion section, incorporating more references into the discussion could enhance its quality.

Technical corrections/suggestions:

• Lines 13-21: Could these paragraphs be combined? Now the first paragraph is very short, which hampers the flow of the text.
• Caption for Fig. 1: The alphabets referring to the subfigures are messed up (“c” is marked twice in the caption, the latter one should be “d” and current “d” should be “e”).
• Line 112: correct “(see A)” to form “(see Appendix A)”
• Line 127: add word “Sect.” to “described in 3.3”. The instructions of TC state: “The abbreviation "Sect." should be used when it appears in running text and should be followed by a number unless it comes at the beginning of a sentence.”
• Line 132: Same here: “(see 3.3)” à “(see Sect. 3.3.)”
• Lines 140-149: Could these paragraphs be combined? Now the first paragraph is very short, which hampers the flow of the text.
• Line 190: Here the “Section” is used. However, correct to: “(see Sect. 3.2)”
• Captions for Figs. 5&6: Be consistent, if you explain the outliers in caption of Fig. 6, explain them also in caption of Fig. 5.
• Fig 6., Table 1, Fig. 9.: To get a more polished outlook replace underscores with spaces. In the caption for Table 1 the abbreviations for study regions are provided but the abbreviations are not used in the table à these could be removed in the name of the consistency (abbreviations are not presented in other captions).
• There are lots of good figures in the manuscript. However, authors could consider combining similar figures together to facilitate comparisons between the research areas. For example, figures 3&4, 7&8, and 10&11 could be combined into subfigures.
• Line 273-273: correct “(described in section 4.4)” to form “(described in Sect. 4.4).
• Line 274: remove extra parentheses after Appendix B.
• First paragraph of the section 5.2.: The paragraph is quite long, and it could be divided into two after the sentence: “In this study, we identify a number of five to six contributions per task as sufficient for assuring a high quality of aggregated crowd-source data.”
• Line 347: Remove “(Volunteered Geographic Information)”. It is unnecessary at this point of the manuscript.
• Last paragraph of the section 5.2.: Quite long paragraph again, which could be divided into two to enhance the readability of it. Division could be made for example after sentence “It has been demonstrated that these types of polygons can be well-represented by Voronoi tessellations (Cresto Aleina et al., 2013; Ulrich et al., 2014).”
• Appendix A: Short paragraphs hamper the flow of the text. All paragraphs could be combined into one.
• Appendix B: Very small detail, but in general the results are presented first for the CB and then to BH study region throughout the manuscript. However, in Appendix B, results are presented first for the BH and then for CB. It would be more consistent to keep the same order if there is no specific reason to change it.

General comments:

Thanks to Walz et al. for the manuscript on evaluating the contributions from non-expert volunteers for the task of mapping ice-wedge polygons in two typical regions in the Arctic. The manuscript is well written and clearly presents the results with tables and figures, demonstrating the potential of using volunteered geographic information to monitor permafrost-thaw features from remote sensing imagery. I have a few comments and suggestions for the authors to consider.

Although technical details of the crow-sourced mapping application are available in Appendix A, however, if some details and potential issues could be mentioned or discussed in the manuscript, it would benefit other crow-sourced applications. A fundamental assumption of using a crow-sourced system is that it allows many people to contribute to a task that cannot be completed by a few experts. If the task does attract many people, how many concurrent users are allowed in the system? How to synchronize data? In fact, it’s not easy to recruit volunteers for mapping specific features (e.g., permafrost-thaw features) that most people are not familiar with, as demonstrated by Huang et al. 2023 (Huang, L., et al. Identifying active retrogressive thaw slumps from ArcticDEM. ISPRS Journal of Photogrammetry and Remote Sensing, 205, 301–316), how are you going to recruit contributors for a continuous monitoring task besides mapping events? How to manage large datasets if you use the crow-sourced system for much larger areas?

Specific comments:

L6: “positional accuracies”, validated against what data?

L13: “the largest non-seasonal component of the cryosphere”? largest in area?                               

A screenshot of the web-based crow-sourced mapping application would be helpful for readers to understand its functions and capabilities.

L146: “crowd-validated”? I am a little confused, as these results will still need to be validated by the experts?

Figure 3 would be good to show a zoom-in region, like Figure 4.

Figure 7, please show a zoom-in Figure, like Figure 8.

L239: Where is the difference between “manually digitized reference polygons” and “expert-derived polygons”?

L286: “betweenness” a sentence to explain betweenness and its importance would be helpful for readers without a hydrological background.

L324: “the overall time available for the crowd-sourced mapping process”, What’s the time referring to? The event duration?

L369: “especially when high-resolution elevation data is unavailable”? This is confusing. This manuscript still requires high-resolution imagery. From my understanding, the need for spatial resolution is determined by the observing objects, that is, smaller features require higher spatial resolution.

technical corrections:

L169: “()rettel-bach2021quantitative”? 

L309: change “inSAR” to “InSAR”.