the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Aug 2026
Multi-decadal evolution of ice-debris complexes at Muztagh Ata (Eastern Pamir) derived from Corona KH-4A and Pléiades stereo images
Abstract. Ice-debris complexes are glacial-periglacial transitional landforms including rock glaciers that provide important insights into geomorphological evolution and represent substantial yet overlooked water reservoirs. However, their formation and evolution remain poorly constrained. Here, we investigate ice-debris complexes at Muztagh Ata, Eastern Pamir, using historical Corona KH-4A stereo images acquired in 1967 and very-high-resolution Pléiades tri-stereo data acquired in 2013 and 2019 to reconstruct elevation changes over five decades and derive surface velocities for the recent period. Our results reveal a three-zone geomorphological organization comprising a glacier, a glacier-affected, and a periglacial zone. The glacier-affected zone exhibits extensive thermokarst features and substantial surface lowering, locally exceeding 100 m between 1967 and 2019, whereas the periglacial zone is dominated by rock glaciers with maximum surface velocities of up to 6 m/yr. These zones occupy approximately 39.4 km², 6.5 km², and 9.4 km², respectively. Despite their considerably smaller area, the glacier-affected and periglacial zones experienced ice volume losses comparable to those of the glacier zone during 1967–2019. Our observations suggest that multiple evolutionary pathways can operate simultaneously within a single ice-debris complex. Glacier-affected landforms may progressively evolve into rock glaciers, whereas debris-covered glaciers may advance onto or become embedded within pre-existing rock glaciers. This study provides new insights into the evolution of ice-debris complexes and highlights the need to incorporate these landforms into glacier inventories and hydrological models for water resource assessments in High Mountain Asia and other mountainous regions on Earth.
Competing interests: At least one of the (co-)authors is a member of the editorial board of The Cryosphere.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(3459 KB) - Metadata XML
-
Supplement
(2472 KB) - BibTeX
- EndNote
Status: open (until 21 Oct 2026)
- RC1: 'Comment on egusphere-2026-3678', Anonymous Referee #1, 01 Sep 2026 reply
-
RC2: 'Comment on egusphere-2026-3678', W Brian Whalley, 15 Sep 2026
reply
Sun et al. (2026) suggest some interesting interpretations of the glaciers and ‘ice debris complexes’ descending from Muztagh Ata relating to stored ice. These complexes are mapped as the ‘glacier affected zone’ including ‘ice-cored moraines’ and ‘debris-covered dead ice bodies’. They are distinct from the lower altitude ‘periglacial zone’ that includes the topographic features ‘rock glaciers’ that have detectable surface velocities. Traditionally, glacial and periglacial have been considered as separate areas as have glacial and permafrost. Sun et al. mostly amalgamate ‘periglacial’ with ‘permafrost’, although though they are different concepts, and suggest that their study ‘offers new insights into glacier–permafrost interactions in High Mountain Asia and beyond’. This is because ‘their formation and evolution are governed by interactions between long-term ice deformation, sediment supply, permafrost and climatic forcing’.
However, a rather different explanation of glaciers and ice debris complexes is possible. Whalley et al. (1986) showed that massive glacier ice was protected from melt by rock debris surface load (‘missing ice’) giving the topographic form known as ‘rock glaciers’ (Whalley, 2025). This also follows from an extension of the disposition of glaciers and debris shown by the Randolph Glacier Inventory (RGI) referred to by Sun et al. (Fig.1b) such as generally seen in mountain areas. These ‘ice-debris complexes’ are merely the extensions of glaciers with a build-up of debris in their lower reaches that modify their ablation and allow glacier ice to extend further down valley. This is the way in which glaciers behave and surging glaciers may play a part of this debris transfer. That rock glaciers and ice cored moraines are of permafrost origin is a misconception.
Glacier ice in rock glaciers, particularly ‘valley floor’ or ‘tongue-shaped’ rock glaciers, has been clearly demonstrated by several authors, notably Noel Potter’s investigations at Galena Creek Rock Glacier (Potter 1972). Nevertheless, Barsch (1996) in his book ‘Rockglaciers’ denied the existence of glacier ice at Galena Creek and exclusively promoted the ‘permafrost model’. Barsch (p. 214) noted that, ’Therefore, Galena Creek rockglacier has to be accepted as a (normal) multiunit rockglacier, which is probably more a talus than a debris rockglacier’. This ‘permafrost’ viewpoint of interpreting the topographic form has been widely and intensively promoted ever since. Subsequent observations at Galena Creek – showing extracted glacier and geophysical results with a surface meltpool ground truth – all falsify the non-glacier ice core idea of Barsch (1996). Unfortunately, these observations have been ignored by those promoting the permafrost model and the current prevailing view (Janke et al. 2013) still follows this ‘creeping permafrost’ model; that the presence of permafrost permits ground ice to form in sufficient thicknesses for it to flow – creep. Thus, rock glaciers, being permafrost bodies, have nothing to do with glacier ice. Further, that rock glaciers can be used to help map permafrost and its zonation.
The association of rock glaciers and permafrost via the ‘Permafrost Zonation Index’ is mapped in Figure 1 of Sun et al.. Permafrost indicates a temperature field that affects the material properties of slopes in mountain landsystems; bedrock, scree, glaciers and rock glaciers, and is approximately zoned altitudinally. According to Sun et al., there are ‘clearly detectable velocities’ in the lowest rock glaciers so these topographic features, following Barsch, indicate frozen soil water, or perhaps snow, but not glacier ice. This is the opposite to what is expected in mountain landsystems; preserved glacier ice but decreased permafrost influence.
Sun et al. (p. 16) suggest that the evolution of ice-debris complexes, following Monnier and Kinnard (2017), can be understood through three main processes (quoted verbatim):
1. ‘the readvance and superimposition/embedding of debris-free/covered glaciers onto/into pre-existing rock glaciers (the "permafrost school")’.- ‘the continuous derivation of a rock glacier from a debris-covered glacier through the preservation and creep of a massive ice core (the "continuum school" perspective).’
- ‘the transformation of a debris-covered glacier into a rock glacier via the fragmentation of the glacier ice and its mixing with periglacial ice and debris’.
These deserve brief comments.
Model 1 assumes that rock glaciers were present before the glacier ice component ‘arrived’ behind, or on top, of them. The model says nothing of how permafrost rock glaciers became to be ‘pre-existing’, bearing in mind that we are in the lower area of the permafrost temperature field. The idea, presumably, is that there is a requirement to introduce glacier ice into a pre-existing permafrost rock glacier body. The latter is rather unlikely.
Model 2 indicates a ‘continuous derivation’ of a ‘massive ice core’. This statement accepts a glacier ice core model for rock glaciers as indicated at Galena Creek (supra) and shown at Gruben Rock Glacier as part of a glacial-debris landsystem (Whalley 2020). Glaciers, debris-covered glaciers and ice-cored rock glaciers are seen in the Hindu Kush (Whalley, 2024) and Himalaya (Peacey et al. 2026), descending to valley floors well below any permafrost on slopes. Complex lobes with glacier ice cores have been demonstrated in the Andes (Whalley and Marangunic 2025). The rock glacier is the surface topographic feature, the deforming material is a glacier with a debris cover.
Model 3 indicates that fragmentation and ice-permafrost mixing is attributed to Monnier and Kinnard (2017) who state, p 507,
‘This is an alternative to the common and controverted [sic, assumed to mean controversial] model of the glacier ice-cored rock glacier where the evolution of the landform is controlled by the expansion and creep of a massive and continuous core of glacier ice’. This is only controversial if, as Monnier and Kinnard do, follow Barsch’s (1996) permafrost ideas. The fragmentation of glacier ice cores and their mixing with ‘periglacial ice and debris’ has yet to be demonstrated.
Note that all three models operate at low altitudes and hence with dubious permafrost presence. Glacier ice is upstream of all the features and conforms to a rheological model of glacier ice flow with low surface velocities as the core descends and thins. Glacier ice is protected from surface melt by accumulating surface debris, particularly from the valley sides. It is unclear how ‘permafrost creep’ applies to rock glaciers with ‘peak velocities up to 6m/yr’ (p. 18). This velocity is more appropriate to full glacier flow than the creep meant to explain low velocities.
One aspect of the Muztagh Ata landsystems not mentioned by Sun et al., is the distinct absence of ‘talus rock glaciers’ e.g. (Forte et al. 2021) on the slopes, in the ‘periglacial zone’ and distinct from glaciers (Brardinoni et al. 2025). The scree slope from about [38.1380,75.1517] at ~5100m descending west to a point some 500m below shows no sign of flow features – as would be expected if snow melt infiltrated and ‘creeping permafrost’ formed rock glaciers. Similarly, the valley descending some 2.4 km southwest from [38.1431,75.1601] at 5600m altitude is in an indicated permafrost zone but shows neither remnant glaciers nor rock glaciers.
A ’simplexity’ (simple origin for complex landforms) explanation of all the features under discussion is that extensive glacier ice lobes have been preserved by various surface debris loads in a mechanically coherent and continuous glacier rheological framework. ‘Permafrost flow’ has nothing to do with these ‘complexes’. Rather than perform expensive geophysical studies over tens of years all we need to do is examine satellite data and note the further appearance of surface melt-water pools on debris-covered glacier snouts. This satisfactorily allows mapping of such complex units as reservoirs of glacier ice in a rational manner.
Brian Whalley b.whalley@sheffield.ac.uk
References
Barsch, D.: Rockglaciers. Indicators for the present and former geoecology in high mountain environments, Berlin: Springer, https://doi:10.1007/978-3-642-80093-1, 1996
Brardinoni, F., Vivero, S., Barboux, C., Bodin, X., Cicoira, A., Echelard, T., Hu, Y., Jones, N., Lambiel, C., Macdonell, S., Pellet, C., Rouyet, L., Ruiz, L., Schaffer, N., Wehbe, M., and Delaloye R.: 2025. RGIK guidelines for compiling consistent rock glacier inventories. Geomorphology. https://doi.org/10.1016/j.geomorph.2025.110050
Forte, A. P., Villarroel, C. D., and Angillieri, M. Y. E.: Rock glacier and protalus rampart inventory in Las Salinas river basin, Central Andes of Argentina. Cuadernos de Investigación Geográfica. http://doi.org/10.18172/cig.4922, 2021.
Janke, J., Regmi, N., Giardino, J., and Vitek, J.: Rock Glaciers. Treatise on Geomorphology. 8.17:238–273, 2013.
Monnier, S., and Kinnard, C.: Pluri-decadal (1955–2014) evolution of glacier–rock glacier transitional landforms in the central Andes of Chile (30–33 S). Earth Surface Dynamics. 5(3):493–509, https://doi.org/10.5194/esurf-5-493-2017, 2017.
Peacey, M. W., Harrison, S., Irvine‐Fynn, T. D., Hu, Y., Anderson, K., Leng, R., Sun, Z., and Liu, L.: First empirical assessment of ice content from a Himalayan rock glacier. Geophysical Research Letters. 53(17):e2026GL124533, https://doi.org/10.1029/2026GL124533, 2026.
Potter, N.: Ice-cored rock glacier, Galena Creek, northern Absaroka Mountains, Wyoming. Geological Society of America Bulletin. 83(10):3025–3058, dol.org/3010.1130/0016–7606(1972)3083[3025:Irggcn]2.0.co;2, 1972.
Sun, Z., Bolch, T., Falaschi, D., Bhattacharya, A.,and Liu, LMulti-decadal evolution of ice-debris complexes at Muztagh Ata (Eastern Pamir) derived from Corona KH-4A and Pléiades stereo images. EGU Sphere. https://doi.org/10.5194/egusphere-2026-3678, 2026.
Whalley, W. B.: Gruben glacier and rock glacier, Wallis, Switzerland: glacier ice exposures and their interpretation. Geografiska Annaler: Series A, Physical Geography. 102(2):141–161, https://doi.org/110.1080/04353676.04352020.01765578, 2020.
Whalley, W.B.: Glacier–rock glacier interactions in the eastern Hindu Kush, Nuristan, Afghanistan [35.92,71.13] in the period 1976–2019. Geografiska Annaler: Series A, Physical Geography. 102(2):141–161, https://doi.org/10.1080/04353676.2024.2321425, 2024.
Whalley, W. B.: The Identification and Diagnosis of ‘Hidden Ice’in the Mountain Domain. Glacies. 2(3):8, https://doi.org/10.3390/glacies2030008, 2025.
Whalley, W. B., and Marangunic, C.: Landscapes and Landsystems: rock glaciers in the mountain slope domain of South America. Journal of South American Earth Sciences. https://doi.org/10.1016/j.jsames.2025.105759, 2025.
Whalley, W. B., Martin, H. E., and Gellatly, A. F.: The problem of “hidden” ice in glacier mapping. Annals of Glaciology. 8:181–183, 1986.
Citation: https://doi.org/10.5194/egusphere-2026-3678-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 152 | 41 | 11 | 204 | 22 | 11 | 11 |
- HTML: 152
- PDF: 41
- XML: 11
- Total: 204
- Supplement: 22
- BibTeX: 11
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In its present form I cannot recommend the acceptance of this submission.
There are two main reasons for this:
L 77. Please give a more precise geolocation of the summit (which has other names too) of the massif. I suggest [38.2776,75.1165] in a standard decimal latitude-longitude.
At least some of the tongues should be geolocated similarly and GLIMS/Randolph Glacier Inventory code on the relevant areas of the exposed glaciers, especially as you refer to the RGI and you could also include a detail of their mapping for one of the tongues.
Figure 2 is very cramped, the components are difficult to distinguish. This is not helped because two ‘eye movements’ are necessary to find out which complex is which and what the colour coding is. The important area under consideration could be enlarged by omitting ’waste’ space’. There is no reason why ‘Complex 1’ could not be shown across a and b for example. In fact, all the diagrams are cramped and it makes it difficult to see what is going on, especially with downstream velocities, trying to visualise the heat maps against the actual velocity range. Further, you have the light pink = ‘rock glacier without clearly detectable velocities’ but show means and error bars, even in the, admittedly low, but usually what is expected of a rock glacier. They are just low velocity components, and are clearly detectable!
Fig 2 you have to start all over again in working out what is what and we leap to a similar situation for Fig 5. It might be better to re-format the information so all the variations are shown relative to each complex and combine Figs 4 and 6 as summaries. This would make comparison, which is one of the major aspects of the paper, much easier to see.
Fig 7 is a much better summary and is really the one to use for interpretation as the colour coding is given explicitly and excess eye movements are avoided. Yet the zones are not stated on it! It could be improved upon by also showing say the summary variation of horizontal velocity from Fig 6 as there is a lot of wasted space. Identification of outer limits could also do with decimal lat-long [dLL] (to 4 figure decimal) positions. Having 2 figure decimal degrees around the edge is ok for cartography but not for glaciological understanding.
Sun et al make an attempt to explain the debris covered glacier tongues as ‘ice debris complexes’ and as ‘glacial-periglacial transitional landforms’ on the periphery of the Muztagh Ata mountain area. Such glacier-cored features have long been noted in the literature as ‘ablation complexes’ as in Johnson (1974) in the Yukon (near to several surged glaciers). There is no need to give these a new collective name, especially as you then give names to the component parts. Also see Gardner (1978) who, correctly, indicated their glacial origin and that, ‘The Wenkchemna Glacier combines characteristics of ice-cored rock glaciers, ablation complexes, and glaciers, supporting the concept of a transition between these features’. Here, Gardner indicates that the transition is surface topography not of deforming mass. It would also be worth noting previous work on glaciers with debris cover done by Martin Kirkbride.
These early Canadian examples show that the rock glaciers are glacier ice cored and relate to the study of Potter (1972) who definitively showed (and with confirmation in several subsequent papers over the years) that Galena Creek Rock Glacier contained a glacier ice core. This has been the basis of the ‘glacier ice core model’ as advocated by other authors although dispute by the ‘permafrost school’. Note that is, of necessity, a continuum model although Sun et al. seem to omit the glacier ice core or glacier model and instead use (as in l 302) "transformation" or "continuum" processes ….. ‘ thick debris mantle, entered a state of thermal disequilibrium and began a slow transition toward rock glaciers under permafrost conditions. I have no idea how this is to take place!
Not the least how with permafrost most likely to occur at high altitudes, where the glaciers are probably sub-zero until they warm by flow (not to mention surging) as well as in their decent to lower altitudes where full body temperature will become polythermal and then ‘temperate’. Yet you want the reader to believe that these glacier and debris covered tongues become permafrost bodies as they descend!
It looks as if you are trying to avoid the mention of rock glaciers having glacier ice cores by quoting, for example, Anderson et al. and calling it ‘continuum’ rather than glacier ice cored (as your overall term is ‘ice debris complex’). Glacier ice cores have been demonstrated by Potter and confirmed by many observations at Galena Creek RG since then. Clearly, you either do not know the literature or woefully disregard it. I am inclined to the latter. For example, you quote Wee et al 2024 for Gruben RG as having ‘an upper glacier affected zone and a lower periglacial part’. (And here you confuse the issue, as in the whole of the paper, by making ‘periglacial’ equivalent to ‘permafrost’ and thus showing ‘permafrost creep’ and, by implication of that model, not-a-glacier.) Yet Whalley’s 2020 (Geog Annaler) paper clearly showed that in the 1850s Gruben glacier at its LIA extent was a full glacier. This is shown, in Whalley’s Fig 7, that the Fletsch glacier (as it was then called) was a clean glacier from the summit of the Fletschhorn to its snout. This is a Swiss Dufour map, under the auspices of Swisstopo. Their maps through time also show the accretion of a debris cover to the retreating northern part of the glacier to become the rock glacier component. In other words, Gruben rock glacier is a glacie with a debris cove, it is an ice-cored rock glacier. To disregard this is not very scientific …
You quote many papers in support of the permafrost model that use geomorphology together with estimates of kinematics. You do not mention any of the recent papers that show the ‘continuum’ is one of continuum mechanics, namely glacier flow governed by the Glen-Steinemann flow law, from upper glacier basins to accretion of debris covers. This cover inhibits melting and allow continued flow to lower altitudes (away from any influences of permafrost). The paper in Geografiska Annaler on Nauratdur rock glacier (10.1080/04353676.2021.1986304) shows a glacier ice core through to the snout. A recent paper confirms this, although now says that it cannot be called a rock glacier because it contains a glacier ice core. More terminological mischief perhaps.
I now make some further comment on the permafrost aspect.
The all the features discussed are superimposed on a ‘permafrost zonation index’ map (Fig 1) although the ‘periglacial zone’ of Figure 7 has both rock glaciers with clearly detectable and ‘without detectable velocities’. (See note above.) The areas of talus and bedrock are excluded from the periglacial area here but linked directly to presence of permafrost and thus permafrost creep of the rock glacier components. The ‘explanation’ of the rock glaciers is thus one due to some form of ‘permafrost creep’. However, and significantly, none of the talus slopes in the vicinity of the periglacial zone of rock glaciers shows any sign of creep-surface flow characteristics (as would be expected with a ‘talus rock glacier’).
Sun et al. view the ice-debris complexes as glacial-periglacial transitions: glacier to glacier-affected to a periglacial zone and this is the appearance when mapped (e.g. their Figure 7). However, in equating the periglacial zone, ‘dominated by rock glaciers’ they assume that the deformation is by ‘permafrost creep’. Further that ‘Glacier-affected landforms may progressively evolve into rock glaciers, whereas debris-covered glaciers may advance onto or become embedded within pre-existing rock glaciers.’ And that, L 275, ‘The periglacial zone of ice-debris complexes is composed of rock glaciers showing coherent viscous flow patterns, ridge- and-furrow topography, steep frontal slopes and lateral margins (Fig. 2).’ However, the ‘viscous flow’ is related to ‘permafrost creep’ (Lines 327 and 334). Unfortunately, the nature of the creep law is not stated but (Zhou et al. 2022) state that, ’the result of interactions between glacier and permafrost, this transition could be divided into three types as suggested by Monnier and Kinnard (2017): i) glaciers or debris-covered glaciers re-advancing and superimposing/ embedding onto/into older permafrost bodies; ii) continuous derivation from debris-covered glaciers into rock glaciers only by evolution of the surface morphology, together with the creep of massive and continuous core of glaciers ice, and; iii) debris-covered glacier-to-rock glacier transformation through both surface morphology and internal structure evolution, i.e., the transition from continuous glacier ice body into perennially
frozen ice rock mixtures (ice-rich permafrost soils).
It is not clear how this transition of materials properties takes place although it is clearly a different deformation expression to the creep (Glen-Steinemann) law for glacier ice.
As noted above, the ‘permafrost model’ eschews this constitutive expression as it does not hold with the
I beg to differ with your permafrost model for the ‘periglacial’ zones. In my view, and I suspect many glaciologists, you present a good case for glacier ice existing in such glacier complexes under debris ‘all the way down’. So yes, they should be included in estimates of ice – but permafrost has nothing to do with it.