Routine high resolution observation of selected major surface currents from space

Transcription

1 1 Routine high resolution observation of selected major surface currents from space Fabrice Collard(1), Alexis Mouche(1), Celine Danilo(1), Bertrand Chapron(2), Jordi Isern-Fontanet(2), Johnny Johannessen(3) and Bjorn Backeberg(4) (1) BOOST Technologies, 115, rue Claude CHAPPE, Plouzané, France (2) IFREMER, Centre de Brest, Plouzané, France (3) NERSC, Edvard Griegsvei 3a, N-5059 Bergen, Norway (4) Univ. of Cape Town, Rondebosch 7701, South Africa

2 2 Doppler shift contributions Observed Doppler velocities = underlying current + background sea state + sea state perturbated by surface current. First order : only underlying current + background sea state Hypothesis based on Doppler observation compared to HF radar except in area where tidal current is fast changing due to interaction with bathymetry : Second order : sea state perturbated by surface current. Advanced models such as Doprim are needed to take into account modification of wave spectrum by surface current gradients.

3 3 CDOP geophysical model function First presented at ENVISAT Cal-Val review in 2002, published in JGR 2005 using wave mode at 23 incidence angle. Modeled using tilt+breaking largest influence from the largest steepness (typically in equilibrium with the wind stress) First order : only wind dependance empirical law only based on wind speed and direction relative to radar look Neural Network training CDOP_23 = f(wind speed/direction)

4 4 CDOP geophysical model function Now extended to all incidence angles using WSM Doppler grid.

5 5 Simple methodology to remove sea state effects Doppler Anomaly/Velocity (ASAR WM) - Residual Doppler Anomaly (current) Radial Wind Speed (ECMWF) Doppler Anomaly/Velocity predicted (NNT) CDOP_23

6 6 Equatorial Pacific Zone monitoring Radial Velocity (wind effect no removed) Radial Current Velocity (wind effect removed)

7 7 Equatorial Pacific Zone monitoring 2D comparisons

8 8 Equatorial Pacific Zone monitoring 1D comparisons

9 9 Equatorial Pacific Zone monitoring Monitoring the seasonal cycle

10 10 First Breakthrough Help to the characterization of errors in existing ocean circulation models (MERCATOR) or ocean current derived using altimetry/scatterometry (OSCAR). The position of the radial current minima obtained from MERCATOR and OSCAR differs from ASAR and Drifters. MERCATOR does not predict the second minimum. Agreement in the position and amplitude of the minima between ASAR and Drifters

11 11 Supersites for systematic acquisition of ASAR Wide Swath scenes (400km width) Gulf stream (North Carolina) Agulhas current

12 NEW : Doppler Centroid Estimation now provided in all ASAR WSM and WSS products Resolution 10km in azimuth and 5km in range 12 Azimuth Azimuth Resolution : 1.25 sec (about 10km) Range

13 Estimation of Doppler anomaly on WS Doppler grid. Anomaly = measured predicted Compensated non-geophysical sources of anomaly : Antenna misspointing (uniform shift) Instrumental bias in the radial direction (radial discontinuities) Doppler estimator bias caused by azimuthal variation fo backscatter (artificial correlation between doppler and sigma0).

14 Total velocities CDOP correction CDOP Residual velocities 14

15 15 Altimetry derived surface current : 3 days mean

16 16

17 17

18 18

19 19

20 20

21 21 Altimetry derived surface current : 3 days mean

22 22 Nov 5, 2007

23 23 Nov 6, 2007

24 24 Nov 8, 2007

25 25 Nov 11, 2007

26 26 Nov 12, 2007

27 27 Nov 14, 2007

28 28 Nov 15, 2007

29 29 Nov 18, 2007

30 30 Validation exercice using altimetry Sept 13, 2007

31 31 Validation exercice using altimetry Agulhas main stream Mesoscale eddie

32 32 Validation exercice using altimetry Sept 16, 2007

33 33 Validation exercice using altimetry Agulhas main stream Mesoscale eddie

34 Agulhas main stream velocity From Sept to Jan Angle to radial axis estimated using tangent to the main stream 34

35 35 Surface drifter velocity (float d378) from the Agulhas Current QuickTime and a TIFF (Uncompressed) decompressor are needed to see this picture.

36 36 Conclusions Large scale surface current now routinely observed by SAR with quantitative estimations of surface current in the radar look direction. Surface current values validated at 100km scale. Need for high resolution validation dataset for full resolution validation (less than 10km). HF radar is an option but the range is to small even for SAR image mode. CDOP applicable to the C band ATI phase. Concept can possibly be extended to X and L bland. Thanks to ESA for now providing systematic wide swath Doppler grid over Agulhas and gulf stream since summer 2007!