the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Aug 2026
First Assessment of SWOT KaRIn Measurements Over Lake Ice and Overlying Snow
Abstract. The Surface Water and Ocean Topography (SWOT) mission, launched on December 16, 2022, carries a Ka-band Radar Interferometer (KaRIn) and provides high-resolution, wide-swath measurements of Earth's surface waters. During the Cal/Val phase (March 29 to July 09, 2023; local date), near-daily observations were acquired over Teshekpuk Lake, Alaska, providing an opportunity to examine the response of KaRIn measurements to lake ice and overlying snow conditions. Using optical and SAR imagery, reanalysis products, and field observations during the same period, we found that SWOT heights provide reasonable elevations over the fully ice-covered (before melt onset) and open water surfaces. Meanwhile, Ka-band backscatter effectively tracks the evolution of snow conditions, freeze/melt events, surface melt onset, and ice break-up, as well as features such as pressure ridges, the inlet stream, cracks, and fragmented ice. Limitations were also found in the SWOT Level 2 KaRIn High-Rate Pixel Cloud Version D data used, particularly in the quality of height retrievals over melting ice, in areas contaminated by specular ringing, and on surfaces with little to no wind. Complementary simulations with the Snow Microwave Radiative Transfer (SMRT) model confirmed the sensitivity of Ka-band backscatter to evolving snow properties (snow density and grain size). These results highlight SWOT's ability to advance the monitoring of ice-covered lakes and their seasonal dynamics, and to support the development of retrieval algorithms of ice and overlying snow properties in preparation for the launch of future satellite missions providing Ka-band frequency measurements, such as the Copernicus polaR Ice and Snow Topography ALtimeter (CRISTAL) and Sentinel-3 Next Generation Topography (S3NG-TOPO) missions.
- Preprint
(3382 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-4309', Anonymous Referee #1, 18 Oct 2026
-
RC2: 'Comment on egusphere-2026-4309', Anonymous Referee #2, 02 Nov 2026
General comments
This study combines near-daily SWOT observations during the Cal/Val phase with optical and SAR imagery, air temperature data, and field measurements to examine KaRIn responses to lake ice and overlying snow. The dense time series provides a useful opportunity to follow short-term changes and identify conditions that affect height retrievals. The observations are relevant to interpreting SWOT data over seasonally ice-covered lakes. Several aspects of the validation, physical interpretation, and conclusions need clarification.
Major comments
1. Scope of the height validation
The field comparison in Figure 13 covers multiple locations but only one date, May 10, 2023. It therefore evaluates performance under the conditions sampled that day, rather than across the different ice phenology stages. The authors acknowledge the limited field observations, but this distinction should also be maintained in the abstract and conclusions: physically plausible heights and a stable time series are not equivalent to independently verified accuracy.
The correlation coefficient of r = 0.89 describes agreement in spatial variation, but does not quantify height errors. Mean bias and RMSE would make the comparison more informative. If suitable measurements from other dates or independent satellite altimetry are available, they could provide additional validation; otherwise, the conclusions should remain within the scope of the available comparison.
2. Spatial relationship between height anomalies and specular ringing
Section 3.1.4 reports ice surfaces 1–2 m above adjacent open water and also identifies specular ringing in Figure 8. It is unclear whether these height anomalies occur within the areas affected by specular ringing.
This spatial relationship is important for interpreting the anomalies. Please identify the affected locations and compare them with the ringing patterns, ice–water boundaries, and relevant quality flags. Anomalies within areas affected by specular ringing should be distinguished from those occurring elsewhere before attributing them to melting or heterogeneous surfaces. The treatment of quality flags should also be stated.
3. Evidence for the low-wind interpretation in Figure 10
The coincidence of anomalous heights and low backscatter is consistent with reduced height-retrieval quality under weak-return conditions. However, it does not establish low wind as the cause. Wind speed and direction near the SWOT acquisition times, potentially from the ERA5-Land data already used, would help assess this interpretation.
The spatial distribution also warrants attention: are the height anomalies scattered, clustered within low-backscatter areas, concentrated along ice–water boundaries, or arranged in regular bands?
The timing of the supporting imagery matters here. On July 05, the RCM and SWOT acquisitions are approximately 11 hours apart, during which floating ice could have moved. On July 09, they are only about six minutes apart, making that comparison better suited to checking the surface conditions at the anomalous locations.
4. The SMRT conclusions do not fully reflect Figure 14
Figure 14 shows that the response to increasing grain size depends on snow depth. Backscatter decreases for shallow snowpacks, whereas deeper snowpacks show increasing or non-monotonic responses. The curves for different snow depths also cross.
The conclusions that increasing grain size and increasing snow depth reduce backscatter are therefore too general. The discussion already recognizes some of these differences; the conclusions should retain the same qualifications. The sensitivity experiments offer possible physical explanations, but do not by themselves establish which snow property drove the observed changes.
Specific comments
1. Identification of pressure ridges in Figure 9
Some linear features have spatial counterparts in the SWOT and Sentinel-1 images, and the later field photographs support the interpretation at the location within the yellow rectangle. However, the SWOT backscatter image also contains a prominent near-vertical contrast boundary that is not explained.
Please clarify how this feature differs from the identified pressure ridges and mark the corresponding features in the two images more explicitly. The field confirmation should remain specific to the photographed location.
2. Stable heights do not demonstrate a limitation in height retrieval
In Section 3.1.2, spatial variations in backscatter accompanied by relatively stable heights are interpreted as suggesting limitations in height retrieval. Changes in snow wetness or scattering properties need not produce appreciable changes in surface elevation. Without independent evidence that the surface height changed, the different responses of height and backscatter are not sufficient to establish a retrieval problem.
3. Distinguishing surface replacement from changes in ice and snow
The manuscript appropriately recognizes that floating ice can move through the fixed buffers. This distinction should also be maintained when interpreting the time series: a transition from ice to open water at a fixed location may reflect either melt or ice movement. A backscatter change alone cannot distinguish the two.
For the melt and refreezing episodes in Figure 6, please clarify whether independent evidence beyond air temperature and SWOT backscatter supports the interpretation.
4. Implications for monitoring outside the Cal/Val phase
The dense Cal/Val observations are well suited to examining short-term changes, but the resulting process descriptions should be distinguished from monitoring capabilities under the less frequent sampling of the science phase. The conclusions should clarify which findings concern KaRIn response characteristics, which support interpretation of individual acquisitions, and which require dense temporal coverage or ancillary imagery. The additional information provided by SWOT relative to the supporting optical and SAR observations could also be stated more clearly.
5. Validation locations and buffer definitions
The first plotted location in Figure 13 has a substantially higher field elevation than the remaining locations. Its identifier and elevation appear to match Tes-2, which is labelled as tundra in the publicly available field dataset. Please confirm the correspondence between the plotted identifiers and the original sampling locations. If terrestrial locations are included, their surface type should be stated and the lake-only validation statistics reported separately.
Please also clarify which analyses use the 100 m buffers around sites 1–10 and which use the 30 m buffers around the field locations. The current description and Figure 1 caption leave this distinction unclear.
6. Editorial corrections
Line 372: “are longer observed” appears to be missing “no.”
Line 631: “decreases from 0.5 to 1 dB” appears to mean “decreases by approximately 0.5–1 dB”; please check against Figure 14.
“Global Climate Observation System” should be “Global Climate Observing System.”
Citation: https://doi.org/10.5194/egusphere-2026-4309-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 499 | 302 | 89 | 890 | 90 | 82 |
- HTML: 499
- PDF: 302
- XML: 89
- Total: 890
- BibTeX: 90
- EndNote: 82
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Lines 85-87 : « Several sea ice studies using SARAL/AltiKa further confirmed its capability for open water and ice-type discrimination, while highlighting that its penetration is generally limited to near-surface layers (Verron et al., 2021) »
Comment: The statement refers to “several sea ice studies,” but only one reference (Verron et al., 2021) is provided. Could the authors include additional relevant references or revise the wording to better reflect the cited literature?
Lines 371 to 372: « This issue has been resolved in version D, which is used in this study, where such artificial height contrasts between open water (bright) and dark water are longer observed (JPL, 2025b). »
Comment: There appears to be a missing “no” in the sentence “such artificial height contrasts … are longer observed.” I assume the authors mean “are no longer observed.”
Lines 510–515: Averaging SWOT observations within 100 m buffers may reduce local noise, but it could also mask spatial heterogeneity associated with transitions between snow-covered ice, melting ice, and open water. Could the authors justify the choice of buffer size and discuss its potential influence on the observed temporal patterns? Reporting the number of valid observations and a measure of dispersion for each buffer could also help assess the representativeness of the mean values.
Line 569: « However, accuracy cannot be commented on with the available information … »
Comment: Although the authors acknowledge that accuracy cannot be assessed with the available information, it would be useful to discuss the uncertainty information provided in the SWOT products. For example, the PIXC, Raster, and LakeSP products provide different types of uncertainty information, including height-related uncertainty fields. Could the authors clarify why these were not considered and whether they could help characterize the reliability of the reported height measurements?
Fig 13. : Adding error bars for both the in situ and SWOT-derived elevations would help readers assess the differences between the two datasets. The authors should clarify how spatial averaging within the buffers affects the reported variability and distinguish measurement uncertainty from spatial variability. Furthermore, since the correlation coefficient (r = 0.89) alone does not establish agreement, reporting the mean bias and RMSE, if supported by the available data, would help quantify the apparent underestimation of SWOT-derived elevations.
Comment on the conclusion: This may be beyond the scope of the present study, but I think it could be interesting to briefly discuss how climate change, particularly changes in freeze–thaw cycles and lake-ice phenology, could affect the interpretation of SWOT observations. The authors could also briefly mention the potential of SWOT for monitoring these changes over time.