The European Sentinel-3A satellite, launched by ESA in February 2016, followed by Sentinel-3B in April 2018, as a part of the Copernicus programme, is the second satellite to operate a SAR mode altimeter.

SRAL is operated in SAR mode over the whole ocean and promises increased performance w.r.t. conventional altimetry. SCOOP (SAR Altimetry Coastal & Open Ocean Performance) is a project funded under the ESA SEOM (Scientific Exploitation of Operational Missions) Programme Element, to characterise the expected performance of Sentinel-3 SRAL SAR mode altimeter products, and then to develop and evaluate enhancements to the baseline processing scheme in terms of improvements to ocean measurements. Another objective is to develop and evaluate an improved Wet Troposphere correction for Sentinel-3.

We present results from the SCOOP project that demonstrate the excellent performance of SRAL at the coast in terms of measurement precision, with noise in Sea Surface Height 20Hz measurements of less than 5cm to within 5km of the coast.

We then report the development and testing of new processing approaches designed to improve performance close to the coast, including, for L1B to L2:
• Application of zero-padding
• Application of intra-burst Hamming windowing
• Exact beam forming in the azimuthal direction
• Restriction of stack processing to within a specified range of look angles.
• Along-track antenna compensation

And for L1B to L2
• Application of alternative re-trackers for SAR and RDSAR.

Based on the results of this assessment, a second test data set was generated and we present an assessment of the performance of this second Test Data Set generated, and compare it to that of the original Test Data Set.

Regarding the WTC for Sentinel-3A, the correction from the on-board MWR has been assessed by means of comparison with independent data sets such as the GPM Microwave Imager (GMI), Jason-2, Jason-3 and Global Navigation Satellite Systems (GNSS) derived WTC at coastal stations. GNSS-derived path Delay Plus (GPD+) corrections have been derived for S3A. Results indicate good overall performance of S3A MWR and GPD+ WTC improvements over MWR-derived WTC, particularly in coastal and polar regions.

Based on the outcomes of this study we provide recommendations for further developments and implementations in the form of a scientific roadmap.

Satellite radar remote sensing is now a well stablished tool to observe sea surface roughness in earth observation. Both nadir and oblique looking radars are efficient sensors to detect sea surface topography and roughness variability, which are many times associated to underwater dynamics at small scales, sub-mesoscales and mesoscales. However, roughness variability relationships with ocean color variables such as chlorophyll concentration and sea surface temperature (SST) have been lacking thus far, and few examples exist in the literature, which are at most anecdotal. In part, this lack of knowledge owes to the scarce exact synergy between radar observations and ocean color and SST measurements. The problem is also complicated by physical factors such as atmospheric stability and its influence on wind generated waves that are paramount in comprehending sea surface roughness. The new synergy capabilities offered by the Sentinel-3 mission, which at present is the only satellite mission that offers simultaneous measurements of ocean color, SST and mean square slopes, has potential to tackle this long standing oceanography problem: to investigate relationships between roughness, ocean color and SST simultaneous measurements and come up with a comprehensive methodology to use those synergy measurements for a better understanding of near-surface ocean physics.
The SAR mode altimeter on board Sentinel-3 is sensitive to surface roughness manifestations originated by dynamic features of various scales in the ocean (Santos-Ferreira et al., 2018). Here we develop and apply a new methodology for automatic detection of high frequency ocean phenomena based on satellite estimation of mean square slopes in the along-track SRAL record. The mean square slope estimates are based on the dual frequency capabilities of Sentinel-3A and 3B radar altimeters, and benefit from the SAR enhanced spatial resolution along-track (300m). The results of the method are partially validated with sun glitter Ocean and Land Colour Instrument (OLCI) images simultaneous with the SAR altimeter. The results suggest that SRAL enhanced Level-2 altimetry data may be used on a routine basis to detect high-frequency near-surface ocean phenomena such as internal solitary waves (ISWs), sharp ocean fronts, river plumes, film slicks with dynamical origin, etc.
We explore several synergy case studies (SRAL, OLCI, SLSTR) in different ocean regions and time scales, making an attempt to understand the correlations obtained between the satellite measured variables in question (namely, mean square slopes, chlorophyll concentration or turbidity, and SST). An attempt is made to isolate the effect of atmospheric stability from the other two variables, by looking at locations where SST variations are small. We focus our efforts in strong gradients of at least two of the ocean variables in question, and describe several case studies in different ocean regions and at different time scales. One of such regions is the continental shelf in the vicinity of the Amazon River mouth. There, many ocean high-frequency features are present, and possibly interact with each other: ISWs, riverine/ocean fronts (plumes), strong currents (the North Brazilian Current), generation of ocean eddies, etc. By looking at exact Sentinel-3 synergies and other auxiliary synergies we explore these ocean phenomena to learn more about the gradients of ocean variables measured with satellite. All this in a very dynamical continental shelf where recently a large underwater reef system was discovered that extends for more than 1,000 km (Moura et al., 2016).

Fully-focused Synthetic Aperture Radar (FF-SAR) is a novel SAR altimetry data processing technique that accounts for the phase evolution of the targets in the scene, making it possible to focus the complex echoes along the aperture. This process, similar to SAR imaging systems, can potentially reduce the along-track resolution to its theoretical limit equal to L/2, where L is the antenna length.

Our recent study (Feng et al., 2018) presented a preliminary assessment of the FF-SAR mode range data from Cryosat-2 (CS2) altimetry in the Nova Scotian Shelf (NSS). With the pseudo-low resolution mode (PLRM) data as a baseline, CS2 FF-SAR parameters (range, wave height, sigma0, and range-gradient derived geostrophic current Vg) were examined to examine if there are any potential improvements, particularly in the detection of nearshore currents in the NSS. Clearly, we found that 1) the noise reduction (i.e. precision improvement) of the FF-SAR data (i.e. range and range-derived current) is apparent and 2) finer-scale signals are identified in the FF-SAR data.

However, the uneven time-space coverage in CS2 SAR mode data in the region limits its further application. As a natural continue to the work, we propose to apply the established evaluation strategy to the SAR mode datasets from Sentinel-3A/3B altimeters that by design provide regular time-space coverage in the region. In the further evaluation analysis, the objectives include 1) to explore a more objective cross-shelf length scale to derive along-shelf coastal current Vg, 2) to investigate whether or not geophysical corrections (e.g. tide, DAC, etc. ) should be applied for an improved detection of the NSS coastal currents, and 3) understand what small-scale signals identified in FF-SAR data represent, such as current, internal waves, or just non-geophysical noises. To reach these, past and ongoing in situ hydrographic and current measurements, and modeling datasets in the region will be used to provide expected mean and dynamical circulation features from the shelf break up to the coast.

We present results of sea level changes along the coasts of Western Africa obtained from dedicated satellite altimetry processing in the context of the ESA ‘Climate Change Initiative’ sea level project. High resolution (20-Hz) sea level data from the Jason-1 and Jason-2 missions over a 15-year-long time span (2002-2016) are considered. These data are based on dedicated reprocessing that combines the ALES retracking system of radar waveforms with the improved X-TRACK geophysical corrections at the coast. We estimate sea level trends over the 15-year long period along the Jason-1/Jason-2 tracks covering the study region, from the open ocean to the coastal zone in order to follow the regional evolution of the sea level trends. Compared to the 1-Hz sea level measurements from conventional altimetry, the retracked 20-Hz measurements allow us to retrieve valid sea level data very close to the coast (within 1 to 3 km depending on the satellite track). This allows us to compute interannual sea level trends in the coastal zone where no information (apart from tide gauge where available) exists. The main objective of this study is twofold: (1) provide sea level products in the coastal areas from reprocessed altimetry data and (2) check whether sea level change at the coast is different or not from that reported in open ocean from conventional altimetry. The Western African region was selected because the tide gauge coverage is almost inexistent (expect for a long record available at Dakar). Besides, the coastal zone consists of a narrow, low-lying coastal belt where a large percentage of the population is settled in a few coastal megacities (e.g., Dakar, Conakry, Abidjan, Accra, Lagos, and Douala), and where most economic activities are concentrated. Results show that over the study period, sea level trends near the coast are significantly larger than offshore. In some instances, they approach 5-10 mm/yr. For each track, we discuss the processes that could explain the observed trends at the nearest points from the coast.

A significant part of the World’s population lives in coastal areas, which are affected by coastal sea level rise and extreme events. Therefore, measuring accurately sea level right at the coast is of utmost importance. In this study we consider the coastal stripe within 50 Kilometres from land and we derive a climate induced relative sea level rise (RSL) along selected coastlines. We combine the RSL rates obtained from altimetry corrected for Vertical Land Motion (VLM) with the RSL from tide gauges and obtain the RSL rates in a set of points along the coastline, which are then combined with open ocean trends to produce a continuous surface. The sea level rise is made available in a regional coastal map.

Satellite altimetry corrected for VLM and in-situ tide gauges provide measurements for this analysis. We show that satellite sea surface height data can now be recovered from classical altimetry (CA) with 4 cm accuracy (compared to 2 cm in open ocean), up to 3-4 km to the coast and that tidal correction is responsible for higher RMS differences between altimetry and tide gauge observations. We further suggest that use of Synthetic Aperture Radar/Delay Doppler Altimetry (SAR/DDA) data from the Sentinel-3 and CryoSat missions provides higher performances in coastal areas, both in terms of accuracy (3 cm) and resolution (up to 300 m to the coast). We then investigate the impacts of improved SAR data in the coastal sea level trends and extreme sea level analysis on selected regions worldwide.

First region is the North Eastern Atlantic shelf, where the SAR data coverage is seven year long and several tide gauge stations co-located with GPS stations are available. Data cover two years of Sentinel-3A (from June 2016 to Mai 2018) and almost seven years of CryoSat-2 (from October 2010 to Mai 2018). SAR products are from the ESA GPOD processor, SARvatore, for both CryoSat-2 and Sentinel-3A satellites and from the Marine Sentinel-3 processor. The coastal dedicated retracker SAMOSA+ is applied in both processors. Reduced SAR altimetry (RDSAR) is from the TALES and STAR in-house products and from the Marine Sentinel-3 product. Model data are taken from the operation BSH model and from the coupled Geestacht COAstal model SysTem GCOAST. In situ sea level and GPS data are from SONEL and BGK/BfG local organisations.

Second region is South-East Asia, where the coverage of SAR altimetry is seven years in Indonesia, but tide gauge stations and GPS stations are rare, Vertical land motion (VLM) is not linear. Results from SAR and from the complete 25-year CA record are compared.

The combination of high tides with extreme sea level events jeopardizes our coasts and our society. Therefore, the ability to model and predict ocean tides is of high importance for coastal management. Despite the latest striking improvements, the coast still remains a challenging zone for tide modeling. Here, the combination of several factors such as complex and poorly-resolved bathymetry, shelf seas, or the water inflow of rivers, negatively affects the performances in tidal estimation.
Satellite altimetry data play a key role in ocean tidal estimation, as they provide precise information on sea level, which is regularly integrated in most of the modern tide models, dramatically enhancing their performance. Coastal regions are challenging environments for satellite altimetry, however the recent large efforts in these areas brought to the development of several dedicated altimetric products, which improved sea level estimations up to few kilometers from the shore.
The DGFI-TUM is working on an updated version of the Empirical Ocean Tide (EOT) model whose focus is to improve the accuracy of tidal estimation at the coast. EOT is a gridded, semi-empirical model that exploits altimetric observations to compute tidal harmonic constants. The model is based on data coming from different missions, in particular: Topex, Jason-1, Jason-2, Jason-3, ERS-1, ERS-2, Envisat, SARAL, and Sentinel-3. For computing sea level anomalies the latest altimetric products dedicated to coastal applications are used, such as the GPD+ wet tropospheric correction, the FES2014 tide model, the DAC-ERA dynamic atmospheric correction, as well as ranges retracked by ALES. The dataset is then harmonized by a multi-mission crossover analysis. Residual tidal harmonic constants are derived through the least-square harmonic analysis of the sea level anomalies, combined with a Variance Component Estimation (VCE) that weights the contribution of the single missions. The values are computed for a grid of 1/8 degree resolution and are provided together with their formal uncertainties.
In this paper we present the methodology used for EOT model computation, some results for sample regions as well as its performance against in-situ data and in comparison with other modern tide models. The study focuses on regions characterized by a complex coastal geometry and tidal regimes, such as the North Sea or the Indonesian Seas. The model accuracy is assessed in terms of root-mean square errors and median of the absolute differences, computed for the single tidal constituents. Finally, the performances of the model are evaluated with respect to the bathymetry and the distance to the coast.