EUMETSAT Invitation To Tender 14/209556: JASON-CS SAR Mode Sea State Bias Study. Final report

This document represents the final report of a study funded by EUMETSAT about SAR mode Sea State Bias (SSB) for the Sentinel-6/Jason-CS mission. The study comprises a critical review of SSB estimation methods in conventional (low-resolution mode or LRM) altimetry, theoretical considerations about the effect of swell on SAR altimeter waveforms and empirical investigations with Cryosat-2 SAR mode data to detect swell effects in L1B and Level 2 Sea Surface Height (SSH). The report concludes by summarising the basis for the selection and derivation of the SAR altimeter sea state bias correction algorithm and the methods available to calibrate and validate SAR mode SSB corrections. Theoretical considerations based on simple SAR waveform modelling indicate that multipeaked waveforms could occur in the presence of swell, but that effects become clearly detectable only when swell height exceeds 4 meters, which is relatively rare. In the case of the Cryosat-2 data examined in this study, only 2% of samples satisfied this condition. Experimental investigations of Cryosat-2 SAR mode data in different swell conditions produce no consolidated evidence of swell effects. Although anomalous 20Hz waveforms are occasionally observed, no statistically detectable effect of swell is obtained in the overall results for average L1B waveform shapes and L2 1Hz SSH biases and precisions. However, it is stressed that analyses in this study were limited geographically by the availability of Cryosat-2 SAR mode acquisitions over the ocean that could be collocated with Envisat ASAR swell data. It is strongly advised that analyses should be repeated with a broader geographical scope, including data from the central Pacific and the Southern Ocean where high sea state and swell conditions are more prevalent. It is suggested that this could be achieved using Sentinel-3 SRTM and Sentinel-1 L2 swell products, should such data be available. Empirical SSB estimation methods offer the only viable way forward at present to estimate SAR mode SSB. Parametric, non-parametric and hybrid methods are all relevant, noting that hybrid methods may provide more robust estimates in those high sea state and swell conditions that are less densely populated and where effects will be more significant. The development of SAR mode SSB corrections should include additional dependence on sea state development, which would be consistent with the tendency in LRM towards three-parameters SSB models (e.g. Tran et al., 2010b; Pires et al., 2016). The challenges of calibrating and validating SAR mode SSB corrections are the same - i.e. no better, no worse - than for conventional altimetry. For SAR mode altimetry however, P-LRM offer a unique way of calibrating and validating SAR mode SSB against conventional altimetry by providing coincident range measurements that have been shown to be unbiased against conventional LRM. In the case of Sentinel-6/Jason-CS, interleaved SAR mode will deliver true LRM data that make it possible to tie the Jason-CS SAR mode mission to the long-term altimetric data record without the issues linked to the loss of precision seen for SAR burstmode P-LRM.