This manuscript presents a methodology for estimating the volumetric storage of soil water within the active layer of periglacial landscapes using InSAR observations. The method from a prior proof of concept publication is re-introduced and applied to a larger area of undisturbed tundra in the vicinity of Toolik Field Station on the Alaskan North Slope using L-band ALOS data acquired between 2006 and 2010. The InSAR-estimated equivalent water depth estimates are compared with in situ estimates, showing strong overall agreement. Lastly, there is a discussion of sources of interferometric error due to DEM errors and SAR image to DEM pixel misregistration.

Overall, I think that this manuscript presents several relevant contributions. However, I would first recommend several revisions before the manuscript can be considered for publication. I have a detailed, point-by-point list of specific comments below, but these comments generally fall under a few broader thematic points. These points are:

1. The manuscript would benefit from a more precise description of what the Zwater parameter physically corresponds to. From my reading of prior work, the Zwater parameter is the equivalent water depth of all pore-bound water undergoing seasonal freeze/thaw within the active layer. However, the Zwater parameter is variously referred to in slightly different ways throughout the manuscript, which could raise questions from readers as to what exactly Zwater physically corresponds (i.e., is this the equivalent water depth, or the thickness of the active layer that is fully saturated?) I have a few specific suggestions to address this.
   1. First, I would recommend including a physical cartoon or schematic that unambiguously illustrates the Zwater parameter. For example, Figure 2 could be modified by including two additional cartoon cross-sections next to the original soil profiles, with the total porebound water separated from the soil matrix into a single homogenous column of pure water, with the depth of this water column labeled Zwater.
   2. I recommend choosing a standard way to describe Zwater and being consistent throughout the manuscript. At various times Zwater is referred to as ‘saturated active layer soil water storage’, ‘saturated soil water thickness’, ‘soil water equivalent depth’, ‘soil water depth in the saturated active layer’, ‘soil water column that experiences the ice-to-water phase change in the saturated active layer (denoted as zwater)’, ‘active layer water storage’ and ‘saturated active layer soil water column’.
2. As currently written, the section of the manuscript on Zwater estimation and the section on DEM errors feel rather disjoint from each other. I think that the manuscript would benefit from a tighter coupling of these two elements, as well as a more thorough and explicit discussion of the novel advancements made in this manuscript on top of the Chen et al. 2020 proof of concept manuscript.
   1. The manuscript would benefit from a more explicit framing of the novel contributions that this manuscript introduces to the field over the prior Chen et al. 2020 study. The discrete dem error due to pixel misregistration analysis is one such novel contribution. However, I would also highlight any advancements made in the Zwater estimation process introduced.
   2. Currently, the Zwater and DEM error sections read as pretty disjoint from each other; these sections could potentially be submitted separately as two stand-alone papers; or, perhaps more discussion linking them could be included in the revised manuscript. For example, to what degree to observed DEM errors propagate into an effective uncertainty for Zwater estimates? Bridging these two sections together through an uncertainty propagation analysis would be one such way to strengthen the cohesiveness of the manuscript.

Individual comments are organized by section below:

Introduction:

Page 1, line 22: “Whether the carbon held by the active layer soils will be transformed to carbon dioxide or methane (a more powerful greenhouse gas), or whether it will flow towards rivers and lakes as dissolved carbon in groundwater, depends largely

on the wetness or dryness of the active layer (i.e., how much water is stored).” I would recommend including a citation or reference to a few relevant papers that support this statement, as it is a central point that underlines much of the scientific justification for this manuscript.

Page 2 line 37: “Because ice density is less than water density (and thus ice volume is greater than water volume), the land surface subsides as the active layer thaws from winter to summer (Liu et al., 2010)” I would recommend being explicit here and stating that the amount of surface subsidence depends upon the overall volumetric water content of the thawing permafrost, as this further motivates the proposed methodology. One second thought, this may not be necessary to state here, as you later state it on line 44.

Methods:

An important point to raise is that, in addition to assuming stationary thaw conditions from year to year, the interannual stacking method also implicitly assumes no variations in excess ground ice content from year to year. While I think that this is a justifiable assumption, it might be worth explicitly mentioning this, and discussing recent work that has demonstrated that InSAR is sensitive to interannual variations in excess ground ice formation and melting: https://doi.org/10.5194/tc-15-2041-2021 ,  

https://doi.org/10.1029/2023WR035331

Section 2.3:

-What are typical pixel misregistration values for ALOS? They are surely processor dependent, and the InSAR processor used is not explicitly mentioned. However, I imagine they are still relatively small, no more than 1 or 2 pixels in any direction, and usually sub-pixel.

Page 10 line 252: This is a minor point, but the comparison between the InSAR pixel (10-100 m) and field measurement (30x30 cm^2 area plot) is a ‘linear to area’ comparison.

Page 11 line 260: I suggest removing, or restating ‘waviness of the ice-table’, as it is not precise.

Page 11 line 260: “we found that the fitted PDFs stayed mostly the same” This statement is also imprecise. Can you quantify what ‘mostly the same’ means?

Results and Discussion:

Section 3.3:

Page 18 line 387: Why do these three scenes likely exhibit more severe pixel misregistration errors compared to the other scenes?

Can an alternative (or compounding) interpretation for the observed phase difference between east and west facing slopes be due to viewing geometry rather than DEM misregistration? Comparing to a descending path frame track over the same area would shed light on this, and allow the authors to rule out a difference in the projection of downslope deformation (e.g., solifluction) onto the LOS vector vs. a pixel misregistration issue.  

Conclusions:

Page 22, line 431: “InSAR-estimated seasonal surface thaw subsidence measures the amount of water stored in the saturated soil active layer above

permafrost, which can be used to constrain hydrologic models and water mass budgets.” Rather than saying InSAR thaw subsidence measures water storage, I might suggest instead something like ‘is sensitive to’ or ‘is related to’, as this is not a direct measurement of soil water storage, but rather a model-based estimation.

This paper builds on previous work by the same research team (Chen et al., Remote Sensing of Environment, 2020), which utilized satellite radar interferometry (InSAR) to estimate frozen ground properties. The current study extends that work by focusing on total soil water storage within the entire active layer column, an important hydrological property that remains poorly constrained across vast permafrost regions.

Major comments:

The previous work by Chen et al. (2020) established a link between seasonal subsidence and active layer water storage.  The primary advancements in this latest work include presenting water storage as a key result, incorporating InSAR data from an additional satellite frame to extend the study area, and identifying InSAR errors caused by DEM defects. While these contributions are valuable, the paper in its current form does not sufficiently emphasize the fundamental novelty of the methodology. The stacking approach and the scaling of seasonal subsidence by density difference to derive total water content remain the same as the previous work. If these methods are not fundamentally new, I recommend shortening the related methodological section and referring readers to Chen et al. (2020) for more details.

The paper highlights DEM error issues that cause noticeable artifacts in individual interferograms at fine, local scales. While these issues are worth noting, the authors did not correct these errors in their InSAR processing. Moreover, it appears that these errors do not affect the final water storage estimates after stacking multiple interferograms (e.g., Figure 5c).

For the HESS readership, I suggest strengthening and elaborating more on the hydrological significance of this method and results. For instance, what are the advantages and limitations of estimating soil water content using this method? What new insights are gained from the estimated water storage in the context of permafrost hydrology?

The rigor of this work could be further improved by providing a quantification of the uncertainties associated with the water storage estimates and discussing the limitations of the method (again, in the context of permafrost hydrology). For example, the simple scaling does not account for unfrozen water, excess ice, or vertical moisture migration; the stacking-based InSAR processing might not be applicable in cases of significant inter-annual variations and linear secular trends; other InSAR errors such as tropospheric delay or phase artifacts due to soil moisture changes are not explicitly accounted, etc.

Minor comments:

Section 1:

Some papers published in recent years have made similar attempts to estimate soil water content above permafrost using InSAR. Consider citing some of them and highlighting your contributions.

• Chen, J., Wu, T., Liu, L., Gong, W., Zwieback, S., Zou, D., Zhu, X., Hu, G., Du, E., Wu, X., Li, R., and Yang S. (2022), Increased water content in the active layer revealed by regional-scale InSAR and independent component analysis on the central Qinghai-Tibet Plateau, Geophysical Research Letters, 49, e2021GL097586, https://doi.org/10.1029/2021GL097586.
• Chen, R. H., Michaelides, R. J., Zhao, Y., Huang, L., Wig, E., Sullivan, T. D., Parsekian, A. D., Zebker, H. A., Moghaddam, M., and Schaefer, K. M. (2023), Permafrost Dynamics Observatory (PDO): 2. Joint Retrieval of Permafrost Active Layer Thickness and Soil Moisture From L-Band InSAR and P-Band PolSAR, Earth and Space Science, 10, e2022EA002453, https://doi.org/https://doi.org/10.1029/2022EA002453
• Widhalm et al. InSAR-derived seasonal subsidence reflects spatial soil moisture patterns in Arctic lowland permafrost regions, https://egusphere.copernicus.org/preprints/2024/egusphere-2024-2356 (Paper accepted for publication in The Cryosphere, title to be changed)

And two recent review papers for your reference.

• Zwieback, S., Liu, L., Rouyet, L., Short, N., and Strozzi, T. (2024), Advances in InSAR Analysis of Permafrost Terrain, Permafrost and Periglacial Processes, https://doi.org/10.1002/ppp.2248.
• Streletskiy, D., Maslakov, A., Grosse, G., Shiklomanov, N., Farquharson, L., Zwieback, S., Iwahana, I., Bartsch, A., Liu, L., Strozzi, T., Lee, H., and Debolskiy, M. (2025), Thawing permafrost is subsiding in the Northern Hemisphere–review and perspectives, Environmental Research Letters, 20, 013006, https://doi.org/10.1088/1748-9326/ada2ff.

Section 2:

(Also relevant to section 3.2) Consider using alternative DEM such as the Copernicus GLO-30 Digital Elevation Model or the latest release of ArcticDEM (v4.1) to quantify and even reduce DEM errors.

Line 153: could you specify the masking thresholds?

Line 163-179: much of its content doesn’t fit within the InSAR processing section and could be relocated.

Section 3:

Figure 5: panel b DEM colorbar's annotations are flipped; please also cite the source of the land cover map in the caption.

Figure 6 caption: What is meant by ‘4% vector length’?

Table 1: only need to keep one significant digit after the decimal points, to be consistent with the description in the main text.

I understand the challenges of directly comparing remote sensing estimates with in-situ measurements. I recommend including scatter plots comparing these in the supplementary materials.

Line 308: PALSAR is misspelled.

Figure 9: add vertical and horizonal scales for the DEM profile and add a distance scale to panel (c)

Add distance scales to Figures 10d, 11a, 12a.

Line 428: Since section 3.3 primarily presents and discusses simulations of errors due to 1-2 pixel misregistration in individual interferograms, it is unclear how your approach provides a valuable way to identify and characterize pixel misregistration errors in the final LOS deformation estimates.