Wave measurements from Space using ASAR image and wide swath mode

Updated dec 10, 2006

Author : dr.fab

After a thorough review the existing retrieval methodologies, the most appropriate ones, considering the different ENVISAT ASAR instrument modes (incidence angle configuration, polarization, resolution and acquisition processing (continuous, burst)), have been implemented in a prototype software. A validation exercise has been undertaken to properly assess performances of the retrieval methodology.

Some improvement have been suggested, implemented and tested following the validation exercises. In particular, a new treatment of the non-linear intrinsic SAR imaging mechanism and non wave signatures (inhomogeneities caused by low wind conditions, internal waves, rain, land contamination...) has been developed. Such improvements helped to gain both in processing speed and applicability to all radar configurations and environmental conditions.

Convincing demonstration of the feasibility to generate SAR Level-2 wave products from Image Mode products was also made. Comparison between SAR and Buoys swell Wave Height (WH) have given a standard deviation error of 0.28 m very similar to those obtained for the Wave Mode product. Performances obtained for wavelength retrieval (std. dev. Error of 39 m) and direction of propagation (Std. dev. error of 12 deg.) can also be considered as very satisfactory.

We particularly emphasize the new potential to derive wave field over large coastal region using the wide swath ASAR complex products. This opens new scientific opportunities to study the interaction between swell waves and meso-scale to large scale surface currents from eddies to well known Gulf Stream and Agulhas Currents. This also opens perspective for near real time coastal monitoring (erosion, sediment transport, marine activity planning, pollutant dispersion ...) by direct use of LEVEL2 products and/or their assimilation in physical and biological regional shelf models.

While the potential to retrieve ocean surface wave in coastal regions was certainly anticipated for standard SAR image products, such a use of Wide Swath products was not previously foreseen at the start of the ENVISAT mission. In combination with Wave Mode products, this will in particular help to advance understanding and solving the known weaknesses of sea state models for long wave (swell) propagation prediction over very long distances.

Validation results

The wave inversion validation was done by comparison with independent directional buoy measurement.



Figure 6: (a) Example of swell retrieval near California such as using ENVISAT ASAR Single Look Complex product. (b) Example of extracted SAR wave spectrum. (c) Collocated Wave spectrum such as measured by direction buoy.


Figure 6 bis: (a) Sea surface roughness map of the Figure6 case. (b) Corresponding SAR wind field.

Each SAR spectrum (left) is partitioned and only the swell part is considered. For each swell partition, the associated swell partition is considered in the buoy spectra (inside the blue contour on the buoy spectral plot (right). All integrated parameters are estimated from the spectral area inside the blue contour to ensure a comparison over corresponding swell systems.



Figure 7: (a) Example of swell retrieval near California such as using ENVISAT ASAR Wide Swath Single Look Complex product. (b) Example of extracted SAR wave spectrum. (c) Collocated Wave spectrum such as measured by direction buoy.

Comparison of integrated parameters

The significant wave height of the corresponding partition in SAR and BUOY spectra is presented for both Image mode and Wide swath mode.

The SAR data used to estimate the SAR spectra are centred over the buoy for the SAR Image products. The nearest SAR spectrum in the inversion grid is used for the SAR wide swath products.


Figure 8 : Significant wave height of the swell partitions from the SAR vs. BUOY spectra for image mode (left) and wide swath mode (right).

SAR/buoy comparison is presented in the following table for both image mode (IM) and wide swath mode (WS). Results obtained from the comparison between ESA Wave mode level2 product (WM) after the latest IFREMER reprocessing and non directional buoys (over 2000 collocation points) is given as reference :

 

RMSe WM

RMSe IM

RMSe WS

Bias WM

Bias IM

Bias WS

Hs, m

0.7

0.28

0.88

0.25

0.19

0.09

Tp, s

1.1

0.87

0.94

0.01

0.27

0.76

Φ, rad

N/A

0.2

0.2

N/A

0

-0.07

Normand Breton gulf : mar 9, 2004

Wave inversion in coastal area using a modified version of ESA level2 wave mode algorithm. Comparison with in situ datawell waveriders and pressure sensor array data from a french navy (SHOM) experiment is presented.

Brittany : oct 5, 2005

1rst Wave inversion processed by the prototype software delivered to ESA on May 24 2006 capable to run automatic wave inversion in shallow water for narrow and wide swath complex ASAR images. The color is coding the significant wave height of dominant wave system and the white arrows (length and direction) are coding the dominant wave vector (wavelength and direction). Intensity of the background image is proportional to the local sea surface roughness.

Hurricane Katrina storm swell : aug 28, 2005

Swell wave inversion over Katrina hurricane and corresponding wind inversion. White arrows indicate wave propagation direction (length of arrows is proportional to dominant wave length), significant swell height is colorcoded. One can observe that the storm swell seems to originate from a location 150km to the North East of the eye.

Graphics generated with SARTool

ASAR data : copyright ESA

Download Image gallery Publications Links © 2004 BOOST Technologies