| 15 Dec 2025
New insights into uncertainties in Antarctic elevation change estimates by comparing radar and laser altimetry
Abstract. Satellite radar altimetry has provided continuous observations of Antarctic Ice Sheet (AIS) surface elevation change since 1992. However, uncertainties in radar-derived elevation estimates remain substantial, primarily due to the influence of local surface topography and time-variable signal penetration into snow and firn. The launch of the ICESat-2 laser altimetry mission in late 2018 established a new benchmark for high-accuracy surface elevation measurements, enabling inter-comparison with radar altimetry results and improved assessment of associated uncertainties. In this study, we use the ICESat-2 measurements to evaluate radar altimetry-derived elevation change estimates from CryoSat-2 over the 6 905 000 km2 large and relatively flat interior of the AIS, where topography-related errors are small. We apply a suite of radar-specific correction methods to the CryoSat-2 measurements, including multiple retracking algorithms and empirical corrections for the time-variable surface and volume scattering of the radar signal. We analyse a 5.5-year overlap period between ICESat-2 and CryoSat-2 (April 2019-October 2024) to assess how the different correction methods influence the CryoSat-2 surface elevation change estimates and their uncertainties.
ICESat-2 observations indicate a thickening of 97 ± 4 km3 yr-1, coinciding with several events of excess snowfall during 2019-2024. All CryoSat-2 solutions yield systematically lower thickening trends, with the smallest bias (0.6 ± 1.0 cm yr−1 or 42 km3 yr-1) obtained using the AWI-ICENet1 convolutional neural network retracker. The remaining trend differences correlate with the ICESat-2 trend signal itself. We discuss possible causes of these systematic differences, one of which is the hypothesis that temporal variations in radar signal penetration associated to temporal variations in snow properties continue to induce systematic errors in inferred surface elevation changes. If the mean trend difference here were representative of the entire grounded AIS (12 352 700 km2), it would correspond to an underestimation of AIS volume and mass trends by approximately 74 km3 yr−1 and 28 Gt yr−1, respectively. These results underscore the challenges of using radar altimetry to resolve subtle, long-term trends related to surface mass balance changes, while also demonstrating the potential of combined laser-radar altimetry analysis to reduce uncertainties in AIS volume and mass balance estimates.
This is an interesting and technically detailed paper comparing CryoSat-2 and ICESat-2 elevation change estimates over the Antarctic interior. The authors apply a comprehensive range of retracking and scattering correction methods and make a useful contribution to understanding radar altimetry uncertainties. I have a few comments below that I believe should be addressed before publication.
Major comments
Section 3.1 – Due to their differing spatial samplings, are 1 km model fits for CryoSat-2 LRM data more poorly constrained than for ICESat-2? How does varying the search radius affect the level of agreement between the two missions?
L320 – 'Since we consider the ICESat-2 elevation changes to be the estimates closest to the truth (i.e. the most accurate estimates), we assume them to be reasonable for validating CryoSat-2 results'. This is a strong assumption. ICESat-2 has its own uncertainties: the green laser penetrates the firn to some degree (Studinger et al., 2024), time-varying subsurface scattering biases could produce spurious elevation change signals (Smith et al., 2025), and scattering from windblown snow could introduce further biases. The authors should more explicitly justify this assumption or caveat their conclusions accordingly.
Figure 5, Table 3, Figure 6 – The reported dh/dt trends in Figure 5 all overlap considering their uncertainties. How does that translate to the small uncertainties for the volume changes given in Table 3?
Section 6.5 – This section does a really nice job of investigating the differences with respect to slope and SMB separately, but are there areas which are both high slope and SMB anomaly? The authors could explore how this affects the partitioning of the variance in delta h.
L524 – The authors discuss the almost zero scale factors in the WAIS, but do not address the variability in the scale factors in the non-D basins (0.19 to 0.65 for OCOG10d). Given the hypothesis that fresh snow drives stronger radar penetration, what does this variability reflect?
L543 – The authors point out the possibility that ICESat-2 measurements in 2023 could be biased and affect the trend, but it is quickly dropped. From Figure 6 it looks like this anomaly is largely driving the difference between the volume trends. I feel this deserves more attention, because if ICESat-2 is anomalous or biased during this period the regression analysis, scale factors and offsets are sensitive to this where ICESat-2 is considered the 'truth'. Indeed, if the 2023 anomaly in J''-K represents a snowfall signal (L411), this undermines the hypothesis that the ICESat-2–CryoSat-2 differences are driven by enhanced radar penetration during snowfall events because the firn model amplitude looks to be in between the two altimeter missions.
Section 6.8 – I have several concerns about this section:
Saying this, the authors are right to highlight that challenges remain.
Minor comments
L4 – 'However, uncertainties in radar-derived elevation estimates remain substantial' — I think this is a somewhat unfair framing of the problem. Uncertainties remain with the laser altimetry and indeed the green light (to a lesser degree) also penetrates into the firn pack (Smith et al., 2025; Studinger et al., 2024). And, as the authors mention later, good agreement has been found between the two missions elsewhere already.
L20 – 'If the mean trend difference here were representative of the entire grounded AIS (12 352 700 km2), it would correspond to an underestimation of AIS volume and mass trends by approximately 74 km3 yr−1 and 28 Gt yr−1, respectively'. Again I find this to be an unfair framing – 28 Gt/yr is less than the uncertainty in Antarctic mass balance reported in IMBIE (Otosaka et al., 2023) and of the order of the uncertainty reported by ICESat-2 studies (Smith et al., 2020).
L61 – 'and their temporal evolution can remain with spurious large variations'. This wording could be improved.
L73 – 'as its green laser virtually has no penetration into the firn layer'. See above comments.
L373 – 'CryoSat-2 OCOG10d reveals larger σAD values than ICESat-2 with mean values of 5.9±3.6 cm vs. 3.6±1.7 cm'. These overlap considering their uncertainties so I'm not sure I would agree one is larger than the other.
L374 – 'For both datasets, but particularly for CryoSat-2, the spatial pattern of σAD seems related to the surface slope and roughness (Fig. 1)'. This can be quantified – ICESat-2 may be similarly related, it is difficult to judge based on the figure alone due to the colour scale.
L382 – 'Overall, the OCOG10d values are larger than AWI-ICENet1c with mean values of 4.9 ± 2.6 cm vs. 4.3 ± 2.1 cm'. As above, not sure I would agree with this statement – they are comparable.
L406 – 'We refer to these as "non-D" basins. Across the non-D basins…'. This seems like unnecessary nomenclature which only serves to confuse the reader.
L527 – 'The throughout positive scale…' apologies I found this sentence hard to read.
L536 – 'As the instrument ages, the transmit power usually reduces and the impulse response widens.' — do the authors have evidence for this occurring in CryoSat-2?
L575 – How do the authors' results compare over the same period (2019–2021)?
Table 4 – Worth mentioning explicitly in the caption why there is no correction d value for AWI-ICENet1.
Review criteria
1. Does the paper address relevant scientific questions within the scope of TC?
Yes. The paper addresses uncertainties in Antarctic ice sheet elevation change estimates from radar and laser altimetry, which is within scope.
2. Does the paper present novel concepts, ideas, tools, or data?
Yes. The comparison of multiple retracking and scattering correction approaches validated against ICESat-2 over a 5.5-year overlap period is a novel contribution, as is the regression analysis linking trend errors to the ICESat-2 signal.
3. Are substantial conclusions reached?
The paper reaches useful conclusions about the relative performance of different retracking approaches, but the conclusion that radar altimetry systematically underestimates thickening due to enhanced signal penetration during snowfall is not fully substantiated. The 2023 anomaly mentioned in the ICESat-2 data appears to drive much of the difference, and the GEMB model at the same time does not support the snowfall penetration hypothesis in the key basins. The implications for AIS mass balance in Section 6.8 are overstated and flawed in some aspects.
4. Are the scientific methods and assumptions valid and clearly outlined?
Mostly. The assumption that ICESat-2 represents truth is insufficiently justified given known uncertainties from firn penetration, subsurface scattering biases and windblown snow. The use of an identical 1 km search radius for both missions despite their very different spatial samplings is not addressed.
5. Are the results sufficient to support the interpretations and conclusions?
Partially. The precision and accuracy metrics are well presented, but the trend analysis is dominated by a single anomalous period (2023) that the authors do not fully investigate. The regression analysis in Section 6.5 depends on ICESat-2 being the ‘truth’, which I feel is not sufficiently justified. The volume change uncertainties in Table 3 appear to be inconsistent with the overlapping dh/dt trend uncertainties shown in Figure 5.
6. Is the description of experiments and calculations sufficiently complete and precise to allow their reproduction by fellow scientists?
Yes.
7. Do the authors give proper credit to related work and clearly indicate their own new/original contribution?
Yes.
8. Does the title clearly reflect the contents of the paper?
Yes.
9. Does the abstract provide a concise and complete summary?
Mostly, though the framing of 28 Gt/yr as an underestimation is misleading without context, and the abstract does not acknowledge the uncertainties associated with treating ICESat-2 as truth.
10. Is the overall presentation well structured and clear?
Yes. The text is well structured and figures are of high quality.
11. Is the language fluent and precise?
There are several areas where the text could be more precise (see comments).
12. Are mathematical formulae, symbols, abbreviations, and units correctly defined and used?
Yes.
13. Should any parts of the paper be clarified, reduced, combined, or eliminated?
Section 6.8 requires revision. Several of the assumptions made there are not physically justified, and the implications for AIS mass balance are overstated relative to existing uncertainty estimates in the literature. For me the "non-D" terminology at L406 adds unnecessary complexity to the reader.
14. Are the number and quality of references appropriate?
Yes.
15. Is the amount and quality of supplementary material appropriate?
Yes.
References
Otosaka, I. N., Shepherd, A., Ivins, E. R., Schlegel, N.-J., Amory, C., van den Broeke, M. R., Horwath, M., Joughin, I., King, M. D., Krinner, G., Nowicki, S., Payne, A. J., Rignot, E., Scambos, T., Simon, K. M., Smith, B. E., Sørensen, L. S., Velicogna, I., Whitehouse, P. L., A, G., Agosta, C., Ahlstrøm, A. P., Blazquez, A., Colgan, W., Engdahl, M. E., Fettweis, X., Forsberg, R., Gallée, H., Gardner, A., Gilbert, L., Gourmelen, N., Groh, A., Gunter, B. C., Harig, C., Helm, V., Khan, S. A., Kittel, C., Konrad, H., Langen, P. L., Lecavalier, B. S., Liang, C.-C., Loomis, B. D., McMillan, M., Melini, D., Mernild, S. H., Mottram, R., Mouginot, J., Nilsson, J., Noël, B., Pattle, M. E., Peltier, W. R., Pie, N., Roca, M., Sasgen, I., Save, H. V., Seo, K.-W., Scheuchl, B., Schrama, E. J. O., Schröder, L., Simonsen, S. B., Slater, T., Spada, G., Sutterley, T. C., Vishwakarma, B. D., van Wessem, J. M., Wiese, D., van der Wal, W., and Wouters, B.: Mass balance of the Greenland and Antarctic ice sheets from 1992 to 2020, Earth System Science Data, 15, 1597–1616, https://doi.org/10.5194/essd-15-1597-2023, 2023.
Smith, B. E., Studinger, M., Sutterley, T., Fair, Z., and Neumann, T.: Understanding biases in ICESat-2 data due to subsurface scattering using Airborne Topographic Mapper waveform data, The Cryosphere, 19, 975–995, https://doi.org/10.5194/tc-19-975-2025, 2025.
Studinger, M., Smith, B. E., Kurtz, N., Petty, A., Sutterley, T., and Tilling, R.: Estimating differential penetration of green (532 nm) laser light over sea ice with NASA’s Airborne Topographic Mapper: observations and models, The Cryosphere, 18, 2625–2652, https://doi.org/10.5194/tc-18-2625-2024, 2024.