Dual frequency airborne SAR experiment on snow water equivalent retrieval in Alpine terrain

Rott, Helmut; Nagler, Thomas; Hetzenecker, Markus; Horn, Ralf; Fischer, Jens; Kubanek, Julia

doi:10.5194/egusphere-2026-2104

Preprints

https://doi.org/10.5194/egusphere-2026-2104

Preprints

08 May 2026

| 08 May 2026

Dual frequency airborne SAR experiment on snow water equivalent retrieval in Alpine terrain

Helmut Rott, Thomas Nagler, Markus Hetzenecker, Ralf Horn, Jens Fischer, and Julia Kubanek

Abstract. In March 2021 a field experiment was conducted in the high-Alpine test site Woergetal of the Austrian Alps, exploring the feasibility and performance of snow water equivalent (SWE) and snow depth retrievals by means of C- and L-band SAR data. Multiple repeat-pass acquisitions were acquired along two flight tracks with an airborne SAR system. The SAR acquisitions spanned two snowfall events of different intensity. In situ data on physical snow properties were collected throughout the campaign. Main objective was the development and evaluation of procedures for SWE monitoring by measuring the interferometric path delay in snow, with focus on the future Radar Observation System for Europe at L-band (ROSE-L), the Sentinel-1 Next Generation mission, and the proposed geosynchronous C-band SAR mission Hydroterra. The response of the VH/VV backscatter ratio and the co-polarized phase differences (CPD) to snow accumulation was also studied. These two parameters do not show a reproducible relation with snow depth and SWE of the snowfall events. CPD measured after snowfall shows an increase with the incidence angle that can be attributed to radar wave propagation through a refrozen crust formed before the snowfall. For the interferometric SWE retrievals three input data configurations were applied: (1) C- and L-band repeat-pass SAR data at full spatial resolution; (2) differential wavenumber (Delta-k) interferograms obtained by split-bandwidth processing; (3) simulated geosynchronous C-band SAR data reflecting the properties of the Hydroterra mission. Coherence, interferometric phase and computed snow accumulation (ΔSWE) images and their properties are documented for these configurations. L-band case 1 shows high coherence and good retrieval performance for both snowfall events. C-band case 1 and case 3 show high coherence and good performance for the low intensity snowfall but low coherence and 2π phase ambiguity prohibit the application for the intense snowfall. This deficiency can be overcome by the Delta-k approach which delivers suitable results for both frequencies. Options for complementary use of the different approaches are discussed.

Received: 13 Apr 2026 – Discussion started: 08 May 2026

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.

Download & links

Preprint (PDF, 2215 KB)

Supplement (1417 KB)

Download & links

Helmut Rott, Thomas Nagler, Markus Hetzenecker, Ralf Horn, Jens Fischer, and Julia Kubanek

Status: final response (author comments only)

CC1:
'Discussion on atmospheric corrections and how this work translates to future EO missions', Benoit Montpetit, 21 May 2026

First, I really appreciated this paper. The methodology and dataset is well presented and described. In terms of results, when looking at other L-Band InSAR SWE retrieval work, the discussion on the coherence is often lacking or completely missing, which I think is a key aspect to enabling the retrieval of SD or SWE using InSAR. This is well presented here, and I really like the discussion on the different challenges between the three different approaches (C-band or L-Band alone, and together). I also like the mention of InSAR using lower frequency overcomes the issue of sensitivity to SWE and microstructure, which is a well known challenge to retrieve SWE using higher frequency backscatter measurements.
One thing I would appreciate if the authors could expand upon is the applicability of these results to current/future EO missions. I am currently investigating NISAR retrievals in Canada with the early released datasets, and I see challenges that were not necessarily addressed in past work, that I feel these results could (not dismissing NISAR data here, I know the team is working on improving it, and looking forward to using it). For instance, satellite L-Band SAR can require atmospheric corrections of several fringe cycles for a single image, and the results shown here present a phase change of 1/4pi between two scenes. This issue is not relevant to this study since atmospheric corrections would be minimal at the given altitude, but how realistic is it to assume that it'll be possible to detect those phase changes when the atmospheric corrections can be several orders of magnitude stronger. How confident is the community on the atmospheric corrections that often rely on modeling. Also, how does scaling up to a space-based platform impact the coherence of the signal. The authors do compare results with changes in resolution, which I think is key when developing an EO mission.
Since the authors already make links with future ESA missions (ROSE-L, Hydroterra), I think it would be nice to push the discussion a bit further with these results, and potentially highlighting challenges for these missions that could be mitigated in the development phases, or expanding upon those challenges that might simply be rapidly mentioned in the conclusion. I know the authors have the knowledge to add a paragraph on how these results translate to the satellite scale.

Thanks

Citation: https://doi.org/10.5194/egusphere-2026-2104-CC1
- AC3: 'Reply on CC1', Helmut Rott, 14 Oct 2026
  
  Publisher’s note: this comment is a copy of AC5 and its content was therefore removed on 15 October 2026.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2104-AC3
- AC5: 'Reply on CC1', Helmut Rott, 14 Oct 2026
  
  Correction of typo:
  We thank Benoit Montpetit very much for the positive encouraging comments on our manuscript. Thanks also for the suggestions to expand upon the relevance and applicability of the study results in respect to the use of satellite based InSAR for SWE monitoring. This is a topic of high relevance considering the current international efforts towards the development and implementation of an operational system for global monitoring of snow mass at high spatial resolution. Regarding the trade-off and complementarity of C-band and L-band, we will address in the Discussion and Conclusions sections in more detail the challenges and options for widespread application of the InSAR SWE retrieval method, taking into account the comments.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2104-AC5
RC1:
'Comment on egusphere-2026-2104', Anonymous Referee #1, 09 Aug 2026

Find comments attached.

Citation: https://doi.org/10.5194/egusphere-2026-2104-RC1
- AC1: 'Reply on RC1', Helmut Rott, 14 Oct 2026
  
  Publisher’s note: this comment is a copy of AC2 and its content was therefore removed on 15 October 2026.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2104-AC1
- AC2:
  'Reply on RC1', Helmut Rott, 14 Oct 2026
  We wish to thank the Referee for the valuable comments and suggestions that are very helpful for improving the manuscript. We address the comments below and explain how these will be taken into account in the revised manuscript.
  Referee comments are in italics, our responses are in normal font.
  General Review: This is a technically strong and generally well-written manuscript presenting a rare dual-frequency airborne SAR experiment with coincident field observations. The comparison of C- and L-band full-resolution interferometry, Delta-k processing, and simulated geosynchronous C-band observations is valuable. The two snowfall events provide a useful contrast between low and high accumulation. The manuscript also gives a balanced negative result for backscatter ratio and co-polarized phase difference as direct indicators of fresh-snow amount.
  I recommend cutting down the manuscript by 20%.
  Response: We appreciate the positive feedback on the scope of our work. This comment reflects very well the main objectives of the paper and the methodologies for snow mass observation considered in the study.
  We will take into account the recommendation for cutting down the length of the manuscript and move information of secondary importance to the Supplement of the paper.
  Major comments
  The manuscript would benefit from a clearer separation between campaign-specific evidence and broader mission-level implications. The demonstrated results apply to two dry-snow accumulation events in three above-treeline RoIs with gentle to moderate terrain. Steep slopes, layover, wet snow, forest cover, and seasonal-scale temporal evolution were not evaluated. Please temper the operational-monitoring claims accordingly and consolidate these limitations in the Discussion and Conclusion.
  
  Response: The properties and constraints of the focus areas for in-situ measurements and retrievals, as well as the temporal constraints are described in Section 2. We will enhance the account on constraints of the SWE retrieval methods in the revised version of the Discussion and Conclusion sections and address open issues regarding satellite-based applications in more detail. The exclusion of wet snow for SAR SWE retrievals, a well-known constraint, is mentioned in line 36. We may bring this up again in the Conclusion.
  Please add a compact performance-summary table that compares all three interferometric configurations for SE1 and SE2. At minimum, report observed and retrieved ΔSWE, bias, MAE/RMSD, coherence, valid-pixel fraction, estimated uncertainty, cell size, and number of independent validation observations. The current presentation relies heavily on three RoI means and makes it difficult to distinguish spatial validation from agreement among regional averages.
  
  Response: Many thanks for this suggestion. We compiled for the revised version of the manuscript a table providing the proposed information.
  A major result of the uncertainty analysis is that reference-phase uncertainty from the corner reflectors is larger than the random phase uncertainty of the snow pixels. This is therefore not a secondary technical detail; it is one of the main limitations of the retrieval. In other words, reference-phase uncertainty appears to dominate the retrieval error, but its estimation from the maximum deviation among a small number of corner reflectors is insufficiently justified. Please provide reflector-specific phase statistics, evaluate sensitivity to the chosen reflector/reference average, and discuss how representative reference phases would be obtained in satellite applications without dedicated corner reflectors or nearby snow-free coherent surfaces.
  
  Response: At first we want to point out that the performance analysis and validation of the SWE and snow depth data derived from the SAR measurements is completely based on in-situ snow data and not on the error estimates provided in Tables 6 and 9. The motivation for the uncertainty estimates in the tables is to point out the need to account not only for random phase errors, but also for systematic errors which in this case refers to the corner reflector phase. We used the maximum absolute deviation of a single reflector as indicator for the corner reflector phase error which is an upper limit. The mean absolute deviation is about 40 % lower. As these numbers are not directly used for uncertainty estimates of the retrieval products, we will move the tables and related text to the Supplement and refer to the computed estimates in the discussion.
  A critical step for ∆SWE estimation is the selection of suitable reference points for phase calibration because the SAR does not measure absolute phase values. This is also the case for InSAR retrievals of surface displacements in geophysical studies for which the reference values are usually stable targets with negligible or known displacements. In case of satellite-based InSAR measurements of SWE the phase of sites with situ measurements of SWE or sites that are snow-free can be used. This is an issue requiring further studies for satellite-based applications. In the current version of the Conclusion section the problem of phase calibration is briefly addressed. In the revised version we will further address this issue.
  The manuscript is well written but excessively long relative to its principal contribution. Please shorten the campaign chronology, detailed backscatter/CPD descriptions, standard methodological derivations, and repeated configuration-by-configuration discussion. The main paper should emphasize the comparative performance and limitations of full-resolution C-band, full-resolution L-band, Delta-k, and simulated geosynchronous processing, with secondary diagnostics transferred to the Supplement.
  
  Response: We will take into account these recommendations, cut down the manuscript and move some information to the Supplement of the paper. Regarding the backscatter/CPD issue, we want to point out that the feasibility and performance of snow depth and SWE retrievals from C-band backscatter intensity and CPD are still topics of dissent. The related account in the manuscript provides full traceability regarding the data base and interpretation based on backscatter intensity and CPD and thus is of relevance for supporting the discussion of this issue. For example, we are not aware of any experimental evidence on the impact of frozen snow crusts regarding the angular dependence of C-band CPD signals. Besides, the evidence of polarization dependent phase differences caused by changes in snow structural properties is also of relevance for InSAR SWE retrievals. Therefore we believe it is appropriate to retain the introduction on the physical basis of CPD changes over snow.
  Minor Comment
  Line 65-70: The discussion about Hydroterra comes out of nowhere. You talk about many other missions and only explain Hydroterra in detail, why? Maybe it can be very brief or go in supplement.
  Response: The Hydroterra mission, an ESA Earth Explorer candidate, was a main motivation for initiating this field experiment. Hydroterra exhibits a new technical concept, namely a C-band SAR operating on a satellite in a geosynchronous orbit. As this concept is innovative and not well known, we include a short technical description in the manuscript. The other SAR missions (Sentinel-1 Next Generation and ROSE-L ) are polar orbiting SAR missions applying the basic technical concept of current operational SAR missions, with various technical advancements. Besides, in the first version of the Introduction the length of the text referring to Hydroterra was only five lines, cut down to four lines in the new version.
  Line 72: What is F-SAR system? Explain before you use it.
  Response: In line 70/71 it is stated that the F-SAR system is an airborne SAR system that operated during the campaign in polarimetric C- and L-band SAR mode. Technical details are specified in Section 2.3 and Table 3. In our view the technical details should rather be retained in Section 2.3 than shifted to the Introduction.
  Overall a careful language and formatting edit is needed to correct typographical errors such as inconsistent “Wörgetal” and “Woergetal”, inconsistent place-name spelling, flight numbering, equation notation, and citation formatting
  Response: Thanks for pointing this out. We will carefully take these issues into account.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2104-AC2
RC2:
'Comment on egusphere-2026-2104', Anonymous Referee #2, 21 Sep 2026

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2104/egusphere-2026-2104-RC2-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2026-2104-RC2
- AC4:
  'Reply on RC2', Helmut Rott, 14 Oct 2026
  We wish to thank the Referee for the valuable comments and suggestions that are very helpful for improving the manuscript. We address the comments below and explain how these will be taken into account in the revised manuscript.
  Referee comments are in italics, our responses are in normal font.
  General comment: This manuscript addresses an important and timely topic, namely the development and evaluation of SAR-based approaches for monitoring changes in snow mass and snow depth in Alpine terrain. Accurate monitoring of these variables is particularly important in Alpine regions, where snow represents a key component of seasonal water storage and where increasing variability in snow accumulation, together with the occurrence of snow-drought conditions, strengthens the need for robust monitoring techniques. However, in its current form, the manuscript is excessively long and, in several sections, overly descriptive. This tends to obscure the main findings and reduces the overall effectiveness of the presentation. I therefore recommend a substantial revision aimed at streamlining and restructuring the text, with particular attention to reducing unnecessary detail and repetition and to highlighting the key scientific messages more clearly.
  Response: We appreciate the positive feedback on the scope of our work underlining the relevance of studies on development and evaluation of tools for monitoring snow mass and snow depth in Alpine terrain. We will comply with the recommendation of streamlining and restructuring the text in order to present the key scientific activities and results more clearly. This will result in a reduced length of the manuscript. Some of the material will be moved to the Supplement of the paper.
  Specific comments:
  The Introduction provides a useful overview of the available radar-based approaches, but the specific knowledge gap addressed by this study could be stated more explicitly. The reader would benefit from a clearer formulation of the main research questions, objectives, or hypotheses at the end of the Introduction. This would also help to better structure the subsequent presentation of the results.
  
  Response: We appreciate the positive feedback on the scope of our work confirming the relevance of the study. Thanks for the advice regarding the revision of the Introduction. We will take into account the suggestions for compiling the revised version.
  Section 2.1 contains a considerable amount of descriptive information, some of which is not sufficiently used later in the interpretation of the results. For example, differences in vegetation cover and surface characteristics among the RoIs are described in detail, but their role in explaining the observed differences between the sites is only partly discussed. Similarly, some information on the evolution of the snow season appears rather detailed relative to its subsequent use in the analysis. I suggest shortening this section and retaining only the information that is directly relevant for interpreting the SAR observations and retrieval performance. The tables could also be used more effectively to summarize site characteristics and avoid repetition in the main text.
  
  Response: Section 2.1 has only 23 text lines, describing main topographic features, the orientation and observation geometry of the radar tracks, the setting of the in-situ measurement points, and surface cover and roughness properties of the different in-situ measurement sites. Furthermore, differences between the individual sites affecting backscatter intensity and InSAR phase are briefly addressed. Consequently, there is not much margin for cutting down this section as the information is of relevance for the analysis and discussion of backscatter properties and of the InSAR phase measurements, as well as for the presentation and discussion of the retrieval results.
  There appears to be a repetition of the word “also” in the sentence referring to C-band changes in snow microstructure and stratigraphy. This sentence should be revised for clarity.
  
  Response: This refers to line 42. We will revise this sentence.
  Lines 86. The study area and the locations of the measurements are introduced before Figure 1 is explicitly referenced. I suggest referring to Figure 1 at this point, as it would help the reader to immediately locate the study area, RoIs, measurement transects, and flight tracks.
  
  Response: We will refer to Figure 1 already at this point.
  Please use a consistent format throughout the manuscript when referring to figures and tables (e.g., ”Figure 1”/”Fig. 1”, and similarly for tables).
  
  Response: We will use “Figure” in the captions and “Fig.” in the text.
  Lines 45-51. Please ensure that the relevant statements are adequately supported by the cited literature.
  
  Response: We will add references.
  The manuscript is strongly focused on the technical performance of the different SAR-based retrieval approaches. While this is clearly the core of the study, the discussion could benefit from a broader perspective on how these results may translate into operational and application-oriented contexts. The manuscript already refers to current and future satellite missions and to some of the associated technical challenges; however, these links could be developed further. In particular, it would be useful to briefly discuss the potential relevance of the proposed retrieval approaches for hydrological applications and water resources management. Given the increasing occurrence of snow-scarcity and snow-drought conditions in Alpine regions, reliable information on snow mass and its temporal evolution is becoming increasingly important for assessing seasonal water availability. Connecting the methodological findings to such broader applications, while also discussing how the identified limitations may affect their transfer to satellite-scale monitoring, would strengthen the overall relevance of the paper and make its contribution more accessible beyond the remote-sensing community.
  
  Response: Many thanks for the recommendations to address topics referring to operational and application-related issues, as well as to point out the relevance of the project results for satellite based applications in snow monitoring. We will take into account these issues in the revision of the Discussion and Conclusions sections in order to consider the project activities and results within broader perspectives. We will describe in more detail also the expected performance and limitations of the different retrieval approaches with regard to operational monitoring tasks.
  
  Citation: https://doi.org/10.5194/egusphere-2026-2104-AC4

Helmut Rott, Thomas Nagler, Markus Hetzenecker, Ralf Horn, Jens Fischer, and Julia Kubanek

Supplement

https://doi.org/10.5194/egusphere-2026-2104-supplement

Helmut Rott, Thomas Nagler, Markus Hetzenecker, Ralf Horn, Jens Fischer, and Julia Kubanek

Viewed

Total article views: 525 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
365	121	39	525	68	34	36

HTML: 365
PDF: 121
XML: 39
Total: 525
Supplement: 68
BibTeX: 34
EndNote: 36

Views and downloads (calculated since 08 May 2026)

Month	HTML	PDF	XML	Total
May 2026	167	47	10	224
Jun 2026	18	5	3	26
Jul 2026	35	14	2	51
Aug 2026	26	10	2	38
Sep 2026	78	30	18	126
Oct 2026	41	15	4	60

Cumulative views and downloads (calculated since 08 May 2026)

Month	HTML	PDF	XML	Total
May 2026	167	47	10	224
Jun 2026	18	5	3	26
Jul 2026	35	14	2	51
Aug 2026	26	10	2	38
Sep 2026	78	30	18	126
Oct 2026	41	15	4	60

Viewed (geographical distribution)

Total article views: 502 (including HTML, PDF, and XML) Thereof 502 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 22 Oct 2026

Short summary

The paper reports on a field experiment in Alpine terrain addressing methods for deriving snow mass from synthetic aperture radar data. Multiple repeat-pass acquisitions spanning snowfall events of different intensity were acquired with an airborne C- and L-band radar system. Three different approaches based on the interferometric path delay in snow were applied for retrieving snow mass, confirming the performance of the method. The complementarity of different processing approaches is shown.


Total:	0
HTML:	0
PDF:	0
XML:	0