Quality Assessment of Surface Current Fields From TerraSAR-X and TanDEM-X Along-Track Interferometry and Doppler Centroid Analysis

Abstract

All existing examples of current measurements by spaceborne synthetic aperture radar (SAR) along-track interferometry (ATI) have suffered from short baselines and corresponding low sensitivities. Theoretically, the best data quality at X band is expected at effective baselines on the order of 30 m, i.e. 30 times as long as the baselines of the divided-antenna modes of TerraSAR-X. In early 2012, we had a first opportunity to obtain data at near-optimum baselines from the TanDEM-X satellite formation. In this paper we analyze two TanDEM-X interferograms acquired over the Pentland Firth (Scotland) with effective alongtrack baselines of 25 m and 40 m. For comparison we consider a TerraSAR-X Dual Receive Antenna mode interferogram with an effective baseline of 1.15 m, as well as velocity fields obtained from Doppler centroid analysis (DCA) of single-antenna data from the same scenes. We show that currents derived from the TanDEM-X interferograms have a residual noise level of 0.1 m/s at an effective resolution of about 33m × 33 m, while DRA-mode data must be averaged over 1000m × 1000 m to reach the same level of accuracy. A comparison with reference currents from a 1 km-resolution tide computation system shows good agreement in all three cases. The DCA-based currents are found to be less accurate than the ATIbased ones, but close to short-baseline ATI results in quality. We conclude that DCA is a considerable alternative to divided-antenna mode ATI, while our TanDEM-X results demonstrate the true potential of the ATI technique at near-optimum baselines.