ENVISAT ASAR level2 wind products

Updated dec 11 2006

Author : dr.fab, V. Kerbaol

Improved Bayesian wind vector retrieval: scientific achievements and recommendations

The most recent and innovative wind retrieval methods have been reviewed to assess their potential performance, their domain of application and validity range. They have been implemented, validated and documented with a particular focus on technical requirements and implications in terms of implementation. The most promising were selected to form wind retrieval prototype software.


Figure 1: Example of graphical map file of generated ENVISAT ASAR level 2 wind field


Figure 2: ENVISAT ASAR level2 wind field mapped in Google Earth

Remarkable progress has been achieved for the wind retrieval since a gain of nearly 20% in the wind speed estimation was obtained compared to the classical algorithms which consists in taking as granted the wind direction given by ancillary a priori wind information. Thanks to an intensive validation against not less than 1000 in situ wind buoys measurements, it has been shown that the software prototype based on an efficient implementation of Bayesian wind vector estimation outperforms the classical inversion scheme for all ranges of wind speeds (Figure 3) of winds direction (Figure 4) and at all incidence angles (Figure 5).


Figure 3: Comparison of the wind speed retrieval performances at various ranges of wind speeds between the Classical inversion scheme i.e. taking wind directions such as given by the ECMWF model and Bayesian inversion implemented in the software prototype.


Figure 4: Comparison of the wind speed retrieval performances at various ranges of wind directions between the Classical inversion scheme i.e. taking wind directions such as given by the ECMWF model and Bayesian inversion implemented in the software prototype.


Figure 5: Comparison of the wind speed retrieval performances at various ranges of incidence anglew between the Classical inversion scheme i.e. taking wind directions such as given by the ECMWF model and Bayesian inversion implemented in the software prototype.

Complementary studies were also conducted to on the specific issue of the wind direction estimation. Encouraging validation results about the ability to detect spectral signatures of wind streaks (RMS of 25° in approximately 50% of cases, see Figure 6) show that there is probably still even room for improvements if one introduce efficiently this additional information as a priori knowledge in the Bayesian estimator


Figure 6 : Histogram of the difference between wind directions retrieved on SAR images using FFT method and wind direction given by in-situ buoy measurements. The RMS error decreases down to 25° if one fit the Gaussian part of the histogram.

Explicit recommendations were also made regarding technical and processing requirements (radiometric accuracy, limitation of noise level, speckle correction and wind cells averaging, use of CMOD-IFR2 rather than CMOD4) and future implementation (use of a priori wind field with known and well-documented uncertainties).

Obviously, the retrieval of high resolution wind fields over coastal areas using Wide Swath mode products stands as the most promising and interesting issue. It provides a unique opportunity to capture the finer details of wind circulation over the largest area with comparable performance in terms of speed and direction estimations at acceptable spatial resolutions (between 500 and 1000 meters). In particular, the capacity to precisely detect and measure wind features that are rapidly varying in space and time (atmospheric fronts, intense and dangerous offshore blowing jets, land/sea breezes, etc) is so far underexploited and should undoubtedly be brought into end-users’ notice.

the remarkable results obtained by the wind retrieval prototype should then incite to a rapid and demonstrative promotion of SAR winds Level2 products. This would as well not fail to stimulate the demand for level3 winds products in domains related to wind energy (Figure 7), marine safety or near real time coastal monitoring.


Figure 7: (left) Map of average wind speeds inferred from SAR on a 1/40 x 1/40° longitude/latitude grid. (right) Map of average wind power in megaWatts assuming air density of 1.225 kg/m3, wind radius of 50m.

Other SAR wind Overlay in Google Earth


kml file (provided google earth is allready installed, right click to save on your disk and double clic on the saved kml file)

Other examples of SAR wind over South Africa Cape Town:

Other examples of SAR wind over Adriatic sea during a Bora wind event:

Other example of SAR wind over Horns Rev Offshore Wind Farm:

High-Res version

High-Res version

 


Graphics generated with SARTool and Google Earth

ASAR data : copyright ESA

Download Image gallery Publications Links © 2004 BOOST Technologies