Thomas Moreau (CLS, France); Claire Maraldi (CNES, France); Marta Alves (CLS, France); Oriane Gassot (CLS, France); Alejandro Egido (ESA - ESTEC, The Netherlands); Laiba Amarouche (CLS, France); Fanny Piras (CLS, France); Pierre Thibaut (CLS, France); Lionel Zawadzki (CNES, France); Francois Boy (CNES, France); Nicolas Picot (CNES, France); Franck Borde (ESA - ESTEC, The Netherlands)

The generation of consistent Essential Climate Variable records from radar altimetry fundamentally depends on waveform retracking, which provides the underlying geophysical parameter estimates. The associated accuracy and precision are therefore strongly driven by the definition and performance of the retracking algorithms. In the context of high demanding requirements in terms of data quality and resolution as of the current reference mission Sentinel-6 and upcoming missions, ensuring statistical optimality and robustness of these methods is now critical.

In this contribution, we present the results of the optimal Weighted Least Square retracking solution, described in Mangilli et al 2025, on Sentinel-6 Ku-band and C-band waveform data, demonstrating that it allows to obtain an unbiased estimation of the geophysical parameters, with significant improvements with respect to current unweighted operational solutions, due to the inclusion of a realistic waveform noise characterization in the estimator. An accurate noise characterisation is particularly important for Sentinel-6 as the pulse-to-pulse correlations linked to the S6 PRF configuration leads to a variation of the Effective Number of Looks (ENL) with respect to both the range gate and the SWH which may impact the parameter estimation if not correctly accounted for.

We provide validation results obtained by applying the WLS analysis to representative Sentinel-6 simulations and to one Ku band and C band data cycle, discussing the impact on the geophysical parameter estimation in terms of bias, uncertainties and parameter correlations with respect to operational solutions. Interestingly, an average range bias of 7.9 cm between the weighted WLS and the operational unweighted solutions is found for the C-band, closely matching the range bias currently applied to the operational solutions. These results are confirmed with simulations. Finally, we present preliminary results of the application of the WLS analysis to 9 kHz and 2 kHz waveform data acquired during the tandem phase of Sentinel-6MF and Sentinel-6B missions. The WLS retracking solution for Sentinel-6 is currently being implemented in the Sentinel Processing Prototype (CNES/CLS). These results demonstrate the benefits of the WLS method across different frequencies and instrumental configurations paving the way for enhanced ocean altimetry with current and future reference missions.