Marcello Passaro (Deutsches Geodätisches Forschungsinstitut. Technical University of Munich, Germany); Jesús Gómez-Enri (Applied Physics Department. Faculty of Marine and Enviromental Sciences, University of Cádiz, Spain); Roberto Mulero-Martínez (Applied Physics Department. Faculty of Marine and Enviromental Sciences, University of Cádiz, Spain); Irene Laiz (Applied Physics Department. Faculty of Marine and Enviromental Sciences, University of Cádiz, Spain); Frithjof Ehlers (Faculty of Civil Engineering and Geosciences. Technical University of Delft, Netherland); Florian Schlembach (Deutsches Geodätisches Forschungsinstitut. Technical University of Munich, Germany); Michele Scagliola (Aresys (Advanced Remote Sensing Systems). Spin-off of Polytechnic University of Milan, Italy)

One of the most novel concepts in coastal altimetry is the Fully Focused SAR (FF SAR) processing. FF SAR introduces improvements in terms of along-track resolution in high Pulse Repetition Frequency (PRF) radar altimeters. The processing is similar to the SAR altimetry, but the along-track resolution up to the theoretical limit equal to half the antenna length (~0.5 m). This is in contrast to the ~300 m unfocused SAR along-track resolution given in coastal altimeter products (Egido and Smith 2017). The FF SAR footprint is pulse-limited across-track and SAR focused along-track. The main drawback of FF SAR is its high computational effort, but some works have been done to reduce the computation time without losing accuracy (Guccione et al., 2018).
The performance of Sentinel-3 FF SAR products in the Gulf of Cadiz (Spain) was analysed and compared with unfocused SAR, in terms of accuracy and precision. The analysis was done in the 0-5 km coastal band, being zero the point where the track intersect with the coast. Two FF SAR algorithms still in development were used: (i) FF SAR Back Projection (BP) (S3 prototype version of Kleinherenbrink et al., 2020); and (ii) FF SAR Omega-Kappa (WK) (Guccione et al., 2018). Four retracking algorithms were used to estimate the retracked range: Threshold retracker (Davis 1993), SAMOSA (Ray et al., 2015), SAMOSA+ (Dinardo et al., 2018) and ALES+ SAR (Passaro et al., 2020). Two tracks from Sentinel-3A and two from Sentinel-3B were processed in the study area according to the availability of in-situ measurements. The analysis was done in the period 2016-2019 for the six datasets.
The products accuracy was obtained by comparing time series of Sea Level Anomaly (SLA) at 80 Hz with those obtained from a radar tide gauge. The statistic used was the standard deviation of the difference (sdd). Also, an analysis of the Percentage of Cycles High Correlated (PCHC) was done. This analysis is done before the outlier detection so the results can be used to compare directly the different dataset. To evaluate the precision: (1) the SLA differences between consecutive measures along-track were calculated for each cycle, (2) the noise over a single cycle was obtained using the average of these differences, (3) the track precision was determined averaging the noise of all the cycles. These differences were considered a good estimation of noise, since SLA is not expected to change significantly in 85 m, which is the distance between consecutive measurements at 80 Hz.
The products still under testing but preliminary results show percentages of PCHC higher in FF SAR than unfocused SAR. The accuracy in FF SAR and unfocused SAR is quite similar. Values of sdd between 5 and 10 cm were obtained in most of the datasets and tracks. Although it should be noted that lower values of sdd were obtained at 1-2 km of the coast in FF SAR. Moreover, preliminary results in terms of along-track precision show lower noise in FF SAR than in unfocused SAR. Differences between FF SAR algorithms and retrackers are analysed too. Some improvements are observed retracking with ALES+ SAR retracker.
In addition, the across-track surface current velocities, derived from the different SLA retrievals, are compared with high-frequency radar (HFR) data, in order to assess the potential of the different algorithms and retrackers for their use in coastal oceanography applications in the Gulf of Cadiz.