Performance Analysis and Assessment of Ocean Current Velocity Measurement Capability for Sentinel-1 Next Generation using dual-channel RADARSAT-2 Modex ScanSAR Data (ATI-S1NGEN Feasibility Study) - Final Report

Kurzfassung

The S1-ATINGEN project was performed in the framework of preparatory studies for the Sentinel-1 Next Generation (NG) C-band satellite Radar mission. This mission is envisioned to have High Resolution Wide Swath (HRWS) mapping capability based upon a multi-channel SAR system. An implementation option for this SAR system is a phased-array antenna with multiple phase centers (i.e. channels) in azimuth direction, allowing for single-platform ATI. This capability enables different ground motion detection and measurement applications, one of those being to measure the radial component of ocean surface velocities caused by circulations and tidal currents. This can alternatively be done also by the Doppler Centroid Anomaly (DCA) approach, which requires just a single SAR antenna phase center (or channel). The main objective of this study was to assess the possible ocean surface velocity measurement performance for Sentinel-1 Next Generation by using an existing C-band, single platform, multi-channel satellite SAR system. Per project definition, ocean surface velocity measurement with the RADARSAT-2 satellite with its MODEX ScanSAR mode were utilized for that purpose. A further objective was to compare the measurement performance of the ATI and the DCA techniques, comprising error, accuracy and sensitivity analysis, both theoretically and empirically. A theoretical analysis of single platform, satellite SAR ATI and DCA ocean surface velocity measurement performance was done generally and specifically for the SAR sensors/missions used in the study. By considering different real measurement scenarios, i.e. different test sites, representative ocean current velocities, current flow directions and sea state conditions, the aim was further to empirically investigate the potential as well as the limits of the ATI and DCA method with RADARSAT-2 MODEX ScanSAR as well as their possible complementary usage. In order to acquire suitable ocean surface velocity measurements with the different SAR sensors, a total of 14 test sites were selected initially for data collection (7 tidal current sites and 7 circulation current sites). In order to compare and to cross-validate the RADARSAT-2 MODEX ScanSAR ATI and DCA ocean surface velocity results with other satellite SAR ATI and DCA measurements, a cross-sensor data acquisition program was performed for the test sites to have spatial and timely overlap of RADARSAT-2 with data collections by Sentinel-1 TOPS (C-band, single-channel, DCA) and by the TerraSAR-X and TanDEM-X missions (X-band, single- and dual-channel, DCA/ATI). Coordination of quasi-simultaneous and collocated data acquisitions with these four SAR missions proofed to be a challenging and elaborate task. The detailed results of the data acquisition campaigns were described in a separate technote. For the processing of the acquired data and for the extraction of radial (RVL) and ground-range projected (GVL) ocean surface velocity products, the project partners newly developed or expanded existing ATI and DCA processing chains for the different sensors. The quality of the pre-processed RADARSAT-2 MODEX ScanSAR and Stripmap data was analyzed as well in a separete technote. The final processing and calibration of these data with a dedicatedly developed processing chain along with the analysis of the resulting ATI and DCA velocity maps were described in a separate technote as well. The processing and calibration of the Sentinel-1 TOPS single-channel data and the analysis of the resulting DCA velocity maps with a dedicatedly developed burst and swath stitching processing chain and the comparative analysis was done and the results described in a technote as well. Processing and calibration of the TerraSAR-X and TanDEM-X data were done with processing chains that were developed or expanded for the needs of the project. This and the analysis of the resulting ATI and DCA velocity maps were described in a technote. In order to conform with the requirement of a common 2-D sampling grid for the extracted velocities and to make results comparable across all sensors later-on, the ATI and DCA velocity maps were processed to a common spatial resolution and grid posting on ground. A cross-sensor validation analysis was performed for the ATI and DCA ocean surface velocity products derived from the measurements with RADARSAT-2, TERRASAR-X, TANDEM-X and Sentinel-1. This comprised a detailed comparison of the collocated and quasi-simultaneously performed ocean surface velocity measurements and an analysis and discussion of the differences observed between the results from the used SAR sensors and between ATI and DCA measurements. These were described in a separate technote. A further objective was (as far as possibl) to validate all ATI and DCA based surface ocean velocity measurements against reference data (ocean current models) or in-situ measurements (e.g. buoy and drifter data). Thus, the validation analysis comprised as well the comparison of the SAR sensor velocity products to ocean surface velocity reference data. This was achieved by using the OSCAR ocean model and, for predicting tidal current velocities, using the FES2014 model. The scope of the study was to focus on the measurement of the directly, i.e. overall observed ocean surface velocities, which are a sum of the radial surface current velocity component, the phase velocity of Bragg-resonant capillary waves and of the long orbital wave velocities. However, as the influence of wind/wave effects on the measured Doppler velocities, be it ATI or DCA, differs for different acquisition geometries (e.g. incidence angle) and for different sensor parameters (e.g. Radar frequency), their influence was taken into account in the cross-sensor validation analysis and in the comparison to reference data in order to explain observed differences.