CoAuthors

Philippe Schaeffer (CLS, France); Yannice Faugère (CLS, France); François-Xavier Davanne (CLS, France); Gérald Dibarboure (CNES, France); Nicolas Picot (CNES, France)

Previous studies underlined the improved accuracy of the up-to-date Mean Sea Surface (MSS) models. They are greatly induced by the use of geodetic altimeter measurements that contributes to reduce the omission errors. Nevertheless, they also underscored residual errors (omission and commission) for wavelengths shorter than ~100km. These errors represent nearly 30% of the SLA variance along the Sentinel-3A tracks and are on the same order of magnitude as the instrumental noise floor of this altimeter (Pujol et al, OSTST 2016). Reducing the MSS models errors at short wavelengths remains thus necessary to fully exploit the altimetry measurement, especially with new and upcoming altimeter technology as Jason-CS/Sentinel-6 or SWOT.
We focus in this study on the interest to use Sentinel-3A measurements for a new MSS model estimation. Collection of nearly 17 months of high resolution measurements now available allows us to compute a first estimation of high resolution MSS along the track of the altimeter (i.e. mean profile). Contrarily to what was observed for LRM altimeters, the low SAR instrumental noise level allows us to obtain a quite smooth mean profile even with few months of measurements.
This mean profile will at first benefit to the Sentinel-3 sea level error budget: its spectral signature comes closer to the theoretical spectrum with a reduction of the MSS error specific signature. This mean profile could contribute to the estimation of a future MSS model, improving missing or erroneously estimated small geophysical structures.