| 24 Jun 2026
Assessing SWOT performance to measure waves and sea level variability over Rangiroa atoll, French Polynesia
Abstract. Atoll reef islands are highly vulnerable to climate change because of their low elevation, exposure to ocean swells, and dependence on coral reef health. However, Sea Surface Height (SSH) and Significant Wave Height (SWH) variability across large atolls remain poorly documented due to limited in-situ observations. Satellite altimetry offers strong potential for monitoring these environments. In particular, the Surface Water and Ocean Topography mission (SWOT), through its Ka-band Radar Interferometer (KaRIn), enables two-dimensional observations of SSH and SWH at unprecedented resolution, offering new opportunities to investigate the barrier-reef lagoon dynamics of atoll islands. Yet, the performance of SWOT observations in such complex coastal settings must first be assessed. In this study, we compare KaRIn measurements with wave buoy, tide gauge, and in-situ Global Navigation Satellite System (GNSS) data collected over the Rangiroa atoll in 2025. Results show that KaRIn measures lagoon SWH with a bias of 10–30 cm and a centered root mean square error (CRMSE) of 10–13 cm, outperforming the global wave model currently used to estimate Sea State Bias (SSB) in current SWOT Level-2 products. Recomputing SSB using KaRIn SWH improves SSH agreement with in-situ observations by up to 3 cm. Spatial comparisons with GNSS-derived SSH also show that KaRIn captures strong SSH gradients near the main pass and northern atoll. Finally, improved SSH anomaly fields using time-averaged KaRIn SSH reveal differences of up to 50 cm between lagoon and ocean water levels driven by wave and tidal forcing.
This manuscript presents a valuable and timely assessment of SWOT performance for measuring sea level and wave variability in a large coral atoll environment. The study is well structured, scientifically relevant, and addresses a topic of clear interest for both the SWOT community and researchers working on reef and atoll hydrodynamics. The results regarding the use of KaRIn-derived SWH to improve the sea-state-bias correction and the reconstruction of a local mean SSH field are particularly interesting.
I recommend publication after minor revisions. My comments mainly concern the positioning of the work within the existing coastal altimetry literature, clarification of some methodological aspects and some results, and improvements to the presentation of some figures and interpretations.
Specific comments
Lines 30-31 and introduction
The discussion of coastal altimetry in the Introduction and especially this sentence appear somewhat outdated and do not reflect the substantial progress achieved in the last two decades. Numerous studies have demonstrated the capability of modern coastal altimetry products to retrieve sea-level information much closer to the coastline than was previously possible. For example, the Sea Level CCI project provides sea-level trends and variability estimates very close to the coast in many regions. Other products, un-retracked such as XTRACK, and retracked such as OpenADB-ALES, have been extensively used for many different coastal studies.
I therefore encourage the authors to substantially revise this part of the Introduction and provide a more comprehensive review of developments in coastal altimetry. In addition, land contamination is only one of several sources of degradation affecting coastal altimetry measurements. In environments such as coral reefs, lagoons, and enclosed bays, the assumption of homogeneous backscatter within the radar footprint may not hold even in the absence of land contamination because different water surfaces within the footprint can exhibit very different backscatter characteristics (e.g., due to spatially varying wind conditions).
I was also surprised not to see any reference of recent developments from the ESA Sea State CCI project, especially considering recent studies that have used altimetry to quantify wave attenuation across coral reef environments. Including these developments would provide a more balanced and up-to-date overview of the state of the art while still clearly highlighting the important advances brought by the SWOT mission.
Line 191 and other occurrences
In many instances where „SWOT Science Team (2024)“ or similar references are cited, corresponding peer-reviewed publications are already available. Whenever possible, I encourage the authors to cite the relevant peer-reviewed literature rather than user handbooks, presentations, or technical documentation.
Line 194
I am unsure about the statement that tide-gauge measurements are not affected by the solid Earth tide, pole tide, and load tide. A tide gauge compares sea level to a benchmark attached to the Earth's crust, and the treatment of these effects in altimetry vs tide-gauge comparisons is discussed extensively in the calibration and validation literature.
For example, Section 2.3 of Kleinherenbrink et al. (2018, https://doi.org/10.5194/os-14-187-2018) discusses these corrections in the context of altimetry validation. It may well be that these effects are considered negligible in the present application, but the categorical statement that tide gauges are not affected by them appears too strong. I would therefore encourage the authors either to provide references supporting this approach or to clarify the assumptions underlying the correction strategy.
Section 5.1
I am somewhat puzzled by the very good agreement reported for SWOT SWH. Several users working with coastal SWOT SWH products have identified known issues affecting both Version C and Version D products. In particular, significant across-swath SWH biases have been observed, varying as a function of beta angle, and these effects are currently only partially corrected through external calibration files (which are not yet included in the baseline). In addition, very low SWH conditions in coastal regions often appear as zero values.
The reported performance is remarkably good, and I encourage the authors to discuss these known issues explicitly and to demonstrate that the statistics shown in Figure 4 are not affected by such artefacts. This additional discussion would substantially strengthen the confidence in the conclusions.
Section 5.1
Some of the co-authors are closely involved in the SWOT nadir altimetry reprocessing activities within the ESA Sea State CCI framework. The will know that retracked and cross-calibrated low-rate and high-rate SWOT-nadir SWH products should be available in the region of interest and, importantly, exactly along the section located between the two deployed wave buoys.
Including these dataset as an additional comparison in Figure 4 would provide valuable context.
Line 340 and similar occurrences
The manuscript reports several interesting performance metrics in terms of bias and random error. I encourage the authors to compare these values with those reported in the existing literature on SWOT coastal performance. Providing such context would help readers better assess the significance of the reported results and understand how Rangiroa compares with other validation sites.
Figure 7
I recommend improving both the caption and the description of this figure.
In particular, it is unclear:
In addition, the methodology used to select SWOT pixels and minimize land contamination deserves a more detailed explanation. More generally, the reader is left wondering what conclusions can be drawn from this analysis regarding the minimum distance from shore at which SWOT SSH measurements remain reliable.
Figure 8 and Section 5.3.1
The discussion regarding the influence of temporal sampling is helpful and suggests that differences in averaging periods do not explain the observed discrepancies between the various mean SSH products.
However, if temporal sampling is not the dominant issue, it would be useful to discuss why the CNES/CLS MSS appears to perform better than the SWOT-derived mean SSH in the vicinity of the tide-gauge location (as suggested by the smaller residuals of the green curve near the tide-gauge position in the lower panel of Figure 8).
Figure 9
The figure would benefit from some improvements in presentation. In particular, the color bars in the first row do not appear to include either a variable name or units. Adding these elements would improve readability and facilitate interpretation of the results.
Finally, there are typos and incorrect English expressions in the text. I recommend reviewing the manuscript with an AI-based proofreading tool (or another automatic grammar checker) to identify and correct them. Here are a few examples:
Line 39: “the the trend”
Lines ~83–85: "originating from the North Hemisphere storms" → "originating from the Northern Hemisphere storms"
Figure 10: “alongthe”
Lines ~161–163 "echoes fromEarth surface"→ "echoes from the Earth's surface"
Lines ~188–190: "corrected of atmospheric..."→ "corrected for atmospheric..." (appears multiple times in the manuscript).
Lines ~280–282: "KaRIn SWH shows bias of 0.28 m and 0.10 m"→ "KaRIn SWH shows biases of ...".
Lines ~309–311: "in most atoll lagoon in the world"→ "in most atoll lagoons worldwide"
Line 364: “Avatory”
The same is true for typos in the list of references, for example:
Dutheil et al. (2020): “Andrefouët"→ should be "Andréfouët" maybe
ECMWF (2024) → Retains placeholder text "accessed: YYYY-MM-DD"