| 29 Sep 2026
Review article: Glacier surges observed by Synthetic Aperture Radar
Abstract. Glacier surges are episodes of strongly accelerated glacier flow. The phenomenon is important to study not only for the advancement of geosciences, but also for risk management and human safety. Unfortunately, surge events remain difficult to observe and predict due to their short duration relative to the entire surge cycle, as well as to their spatial heterogeneity and diverse manifestations. Over the past decades, satellite remote sensing has become the primary source of information for monitoring surge activity, with synthetic aperture radar (SAR) offering unique advantages in terms of temporal coverage and independence from illumination and weather conditions. This article provides an overview of SAR-based methods applied to the study of glacier surges, including interferometry, offset tracking, altimetry, backscatter and coherence map analysis. We discuss how individual SAR products detect complementary aspects of surge dynamics, such as velocity increase, elevation change, terminus fluctuations, surface crevassing and the propagation of instability front. Based on the reviewed literature, we identify current research trends ranging from surge detection and glacier classification to long-term monitoring of individual glaciers, and we point to the main challenges that arise from the characteristics of surge behavior. Finally, we discuss directions for future research, including the integration of multi-aspect SAR observations, the use of machine learning techniques, and the potential role of new SAR methods in early detection and prediction of surge events.
I, Erik Schytt Mannerfelt, reviewed the manuscript of Czyżewski and Milczarek who review the use of spaceborne Synthetic Aperture Radar (SAR) products to observe glacier surges. I am signing this with my name because the manuscript cites many papers that I have (co-)authored, and some of my subsequent recommendations below point to these papers (e.g. Lovell et al. 2026; Harcourt et al. 2026). Therefore, I want to present this review with the transparency that the authors deserve.
The manuscript reviews the history of knowledge on surging and ends with the present questions that may be studied with spaceborne radar satellites. The different manifestations of a surge are then explained which sets the scene for characterization of them. Finally, different spaceborne SAR methods are explained, and the discussion chapter covers present challenges and future opportunities.
My take on this manuscript is that a valuable and useful contribution to the scientific literature on the subject is possible after major revision of the current version. The strength is that the text can be a very useful summary for people that are new in the field, and I find many explanations well-written and easy to follow. The major weaknesses, which I believe can be fixed, are as follows:
General comments
Minor comments
Specific comments
L39: I recommend using the 2026 peer-reviewed paper that came out of Harcourt et al. (2025) instead of the SESS report: https://doi.org/10.1016/j.earscirev.2026.105410
L56: Jiskoot et al. (1998) is not very current and the paper also states "less than 1%". Lovell et al. (2026) is an alternative reference with updated numbers that state >1%.
L64: Surges have been documented in the Antarctic Peninsula: see Davison et al. (2026; https://doi.org/10.5194/tc-20-4293-2026) and references therein.
L80: I don’t know what the sentence ending with “they change their cycle” means. Please rephrase.
L93: I also don’t fully understand what the last sentence here means. Is it a segue to the next paragraph? If so, it should be rephrased to be clearer because “what happens between two surges” is very ambiguous.
L102: “the entire system is gradually losing stability” is vague. I presume that you mean that the surge front tends to spread up- or down-glacier. I would rephrase this sentence to talk about the propagation of a fast-flow regime instead because that is the core difference between a full surge versus a fast-flow regime (and a surface bulge) that never manages to spread to the rest of the glacier (i.e. “partial surge”; Sund et al. 2009; https://doi.org/10.3189/172756409789624265).
L114: I would not say that stagnation and “virtually motionless” is proven enough to state. What we know is that sliding/sediment deformation suddenly drops to insignificant levels, and what remains is mostly ice deformation (which may also be lower due to an extended glacier body leading to a lower surface slope). So it is definitely one or several orders of magnitude slower, but it’s not motionless.
L114-118: This paragraph contains only one reference. I presume that definitions and descriptions are taken directly from Jiskoot (2011), but then the paragraph needs to describe this properly at the start. Something like “Jiskoot (2011) describes stagnation as […]”. An alternative is to find more references, which should not be difficult.
L120: I don’t understand the first sentence. A glacier’s surge history is recorded in the Randolph Glacier Inventory since about ten years, and this is the closest to officially recognized as we can get in the glaciology community. Maybe I missed the point, but this needs rephrasing.
L124: I don’t understand what “an individual process” refers to here. That surges look different at different glaciers? Please rephrase.
L125-136: I urge reading Guillet et al. (2025; https://doi.org/10.1017/jog.2025.10065) closely: there are indeed “slow” and “fast” looking surges, but it is not statistically defensible to separate them by region like the older Alaska- vs Svalbard-type surge classification of the past. I sincerely suggest the authors to consider whether they feel it is defensible to carry this regional dichotomy forward.
L150–154: The description is hard to follow. The key transition in the thermal mechanism is the bed switching from frozen to thawed, enabling sliding; "temperate ice covered with cold ice" being "replaced" obscures this.
L226-231: Another case of many statements needing references followed by one reference. Is the reference for local ecosystem changes and adjacent glaciers from Harcourt et al.? It’s very unclear to me.
L278: DInSAR is presented as if it is a solution to phase noise, but that is precisely what DInSAR does NOT address. Some reorganization of the sentences here may solve it.
L359-383, L509-517 and Table 1: The distinction between speckle and intensity tracking is unclear. Both are usually from amplitude cross-correlation and they mainly differ in what is being tracked (speckle or large-scale features like crevasses). "Coherent" and "incoherent" is used both for the algorithm (whether it uses phase) and for the surface (whether speckle is preserved) which should be kept apart. Leinss et al. (2022) is a good reference resource here.
L619-620: The way the sentence starting “Machine-learning and deep-learning” needs a reference unless rewritten. Right now it’s presented as if we already know that this can be used for detection but it has not been proven yet. Rephrasing it as an interesting potential direction would be more appropriate.