7th bimonthly progress report

Whithin this reporting period, two papers have been submitted to IEEE, and one of them is already accepted with minor revisions.

The accepted paper will be published in a special issue of IEEE TGRS for the EUSAR 2010 workshop. This paper describes for the first time in detail the complex calibration procedure necessary to retrieve Doppler velocity from ASAR single antenna Doppler centroid measurements, including a thorough error analysis. The calibration error of the Doppler anomaly with the present algorithm is about 5 Hz, which corresponds to about 20-40 cm/s, depending on incidence angle, when converted to surface velocity (eventually current). This is in other words a critical factor for currents of weak to moderate strength (< 1 m/s), although, as found and described in an earlier proceeding paper, accuracy increases strongly when creating climatologies by averaging over periods of several months.

The second paper submitted to IEEE TGRS is a relatively simple paper which describes the Bayesian SAR wind algorithm which utilizes the same Doppler velocity. The validation shows that the algorithm works very well for complex wind situations, but that the calibration uncertainty of the Doppler is a limitation in the general case, where the Doppler basically adds only noise to the inversion scheme. A follow up paper is planned, where more details and variations of this algorithm is studied, with a more massive validation against buoys and ASCAT scatterometer wind measurements.

An overview of the presentations and publications within this project, including the above mentioned, is available under "Dissemination" from the menu to the left.

A project "Sentinel-1 bistatic feasibility study", funded by ESA, which complements WP 2 of IncuSAR, has been finalized in this reporting period. The consortium consisted of Johnny A Johannessen and Knut-Frode Dagestad from NERSC, and Harald Johnsen and Geir Engen from Norut IT in Tromsø, Norway. The final results of this study were presented at ESTEC 14 December 2010, and the project final report is soon to be completed. The Statement of Work is available at the bottom of this page, and the final report wil be added as soon as it is completed (end of January 2011).

The purpose of this complementary project was to investigate the feasibility of having a second passive slave satellite flying behind Sentinel-1, for retrieval of ocean surface currents with the along track interferometry (ATI) technique. The phase derived from the ATI measurements can be used to calculate the Doppler velocity, which is the same quantity that is retrieved from ASAR today by another technique. From two satellites and the ATI technique, one can however expect to retrieve the Doppler velocity at a much higher spatial resolution, on the order of hundreds of metres, and to avoid the very complicated post-calibration procedure which is necessary for the ASAR single antenna Doppler anomalies.

For this study, Doppler velocities were simulated with the M4S model developed by Roland Romeiser from the University of Hamburg, Germany. This model was chosen because it contains several SAR-specific features (e.g. antenna noise level and phase retrieval) with are not included in the DopRIM model. A work package directly overlapping WP2 of the IncuSAR project, consisted in inverting back the surface current from the modelled Doppler velocity field. The method tested for the modeled field will in the next period be applied and validated on real ASAR Doppler velocity measurements. The principle of the method is nearly trivial:

The first guess surface current estimated from ASAR Doppler velocities is obtained by subtracting the wind contribution to the Doppler by inputing model (numerical forecast/analysis) wind to the CDOP empirical function. This first guess current does not account for the fact that the current itself modifies the relative velocity of the wind and the moving ocean surface. A second iteration consists of inputing to CDOP the same model wind, where the first guess current velocity is subtracted. The model study indicates that this method may give improved results, but also the this simple method cannot compensate for the very complex intricacies which result from wind-wave-current interaction.

Another paper is currently in praparation, to be submitted to a special issue of Remote Sensing of the Environment with a deadline of 25 January 2011. In this paper the DopRIM model wil be used to model the Doppler and NRCS for strong tidal current fields, where both numerical current fields and SAR observations exist. Based on this, the possibility for 2D current inversion will be further studied. For this paper, Doppler shift will not be obtained from ASAR, but from Single Look Complex Image Mode scenes ordered through EOLI. This will give some more freedom to adjust the spatial resoluton, as a trade-off against accuracy.