Skip navigation.
Home
INCUSAR
Home

5th bimonthly progress report

Work has been ongoing together with Morten W. Hansen at NERSC on the main challenge of improving the calibration of the Doppler anomaly from Envisat ASAR, i.e. the challenges of correcting for:

  • relative satellite-earth motion (which is very large, order of 300 m/s, and therefore has to be calculated very accurately)
  • artefacts due to azimuth gradients of backscatter (sigma0) within Doppler estimator area (approximately 4x10 km)
  • variation of Doppler along range (look) direction, where artefact from antenna pattern is dominating due to poorly known antenna pointing direction 

The latter artefact shows signatures of the antenna patterns very clearly in the Doppler anomaly, but the challenge is that it varies in time in both shape and absolute level (shift). One tested method was to determine the shape of this part from Doppler measurements over the Amazonas rainforest, which has uniform backscatter and terrain elevation, and no wave motion induced Doppler. This absolute level (shift) has then been determined by requiring that the average Doppler over land for scenes before and after a given ocean scene should be zero. For this estimator to be stable, only land with terrain elevation below 500 m should be used. Gradually, by making small improvements in the code, the Doppler velocity seems to improve. The Doppler anomalies calibrated as described above, have been converted to Doppler surface velocity, and the wind contribution has been subtracted with the CDOP empirical function, taking NCEP wind as input.

The remainder, ideally the surface current, has been compared with a moored current meter located at 100 m depth at 60.8 North, 4.25 East, off the North-Western Norwegian coast. In addition we have also compared with surface currents from altimetry, based on an updated Mean Dynamic Topography (Rio et al. 2009). Alle current measurements are projected into the line of sight of the ASAR, which is close to the main current flow direction at this site. Only ascending passes have been used for this comparison. All three datasets are shown on the figure below, plotted versus time from July 2007 until January 2010. The measurements have been smoothered with a 90-day filter; the standard deviation is shown with the blue shaded area for SAR and the red shaded area for the mooring.

It is seen that the smoothed wind corrected ASAR Doppler velocity compares generally well with both the moored current meter and the altimeter current, with an average speed (component) around 30 cm/s. We see that the SAR current is higher than the other two measurements in winter (in particular 2008-2009), when winds are higher. The green line shows 2% of the radial wind, which as a rule of thumb corresponds to the wind driven (Ekman) current. Investigation is needed to find if this is an error of the SAR Doppler method (i.e. wrong CDOP wind removal), or if the difference is a result of different imaging methodologies (i.e. the mooring beeing located at 100 meter depth). The standard deviation (blue shaded area) of the SAR velocity is still quite high, so the uncertainty of the Doppler for a single scene is estimated to be around 20-30 cm/s, comparable to the mean current at this site.

Colleagues at Norut IT in Tromsø, Norway have developed a SAR Doppler estimator which can possible improve the Doppler anomalies compared to the standard ESA WSM estimator. We have ordered and downloaded 20 ASAR WSM Level0 scenes covering the same area, and will test this estimator as soon as Norut can provide a ready version of this code.

 

In parallell to activities described above, ongoing work covers:

  • For WP2 a 2d current inversion scheme will be developed, based on the DopRIM model. This software (developed by Vladimir Kudryavtsev) is quite complex, and with no written documentation, except for published journal papers. It is thus more a research code than a simple ready-to-use software. Some time is therefore needed to get aquainted with the code, as we have already experienced that incomplete understanding and inaccurate specification of any of the many input parameters can yield erroneous results.
  • The Bayesian sceheme developed earlier needs more validation before it can be published as a journal paper. More validation and testing will be performed to be presented at the Living Planet symposium in Bergen June/July 2010.
  • In May 2010 Knut-Frode Dagestad attended the ESA-MOST Dragon2 synposium in Guilin, China. For this project surface current signatures of oceanic internal waves have been studied, and also modelled with DopRIM. This work is complementary to the main IncuSAR tasks, and the results will be directly useful for WP2. This work will also be presented at the upcoming Living Planet symposium.

 

Updated list of conferences attended:

  • SeaSAR 2010 workshop, ESRIN 25-29 January 2010, 1 talk given and 1 proceeding paper submitted
  • Dragon2 Workshop, Guilin, China, 17-21 May 2010, 1 talk given and 1 poster presented

In addition, results and plans of IncuSAR project has been presented at several small local (and non-local) workshops/meetings.

Future planned conferences with talks and proceeding papers:

  • ESA Living Planet symposium, Bergen, Norway, 28 June to 2 July 2010, 2 talks will be given and one proceeding paper will be written.
    • Additionally 2 talks/papers will here be given by Morten W. Hansen, where Knut-Frode Dagestad is second author.

Journal papers in preparation:

  • A SAR Bayesian wind retrieval scheme with validation, based on work presented at SeaSAR 2010
  • Technical journal paper about calibration of Doppler velocity
  • Monitoring the North Atlantic and Norwegian Sea current with ASAR Doppler velocity, based on work presented by Morten W. Hansen at SeaSAR 2010
  • A SAR transfer function for oceanic internal waves, based on work presented at Dragon2 workshop (and upcoming Living Planet Symposium)
  • Planned final main paper (yet not started): An inversion scheme for 2D ocean surface current based on SAR

 

Reference: M.-H. Rio, P. Schaeffer, G. Moreaux, S. Bourgogne, J.-M. Lemoine, and E. Bronner. A New Mean Dynamic Topography Computed Over the Global Ocean from GRACE Data, Altimetry and In-situ Measurements. In Poster communication at OceanObs09 symposium, 2009.