Validation of ERS-2 SAR offshore wind-speed maps in the North Sea

Abstract

Wind maps are retrieved from ERS-2 Synthetic Aperture Radar (SAR) scenes by the CMOD-IFR2 and CMOD4 algorithms for 61 cases at the Horns Rev site in the North Sea and compared to meteorological in situ observations from a mast located 14 km offshore. The in situ data are corrected for flow distortion and sea-level changes prior to validating the SAR wind maps. The SAR wind maps are area-averaged by a simple footprint method assuming neutral stability and with three nonlinear weighting footprint methods including correction for stability. From a physical point of view, the latter is more correct. However, between in situ and SAR-derived wind-speed estimates comparison results of the nonlinear footprint values are statistically less correlated (R ²=0.73–0.77) and the standard error (SE) is larger (>1.5 m s⁻¹) than results from the simple footprint (R ²=0.78–0.80 and SE=1.3 m s⁻¹). The results are found with wind direction determined from wind streaks in the SAR images by Fast Fourier Transform. Using in situ wind direction as input to the CMOD-IFR2 and CMOD4 algorithms yields even better linear regression results, e.g. for the simple footprint method R ²=0.88 and SE=0.9 m s⁻¹. SAR wind maps may be useful for mapping of future offshore wind resources.

Acknowledgments

We are grateful to (1) the European Commission for the WEMSAR contract ERK6-CT1999-00017; (2) ESA for the AO3-153 and EO-1356 projects granting the ERS-2 SAR scenes; (3) Elsam Engineering for the meteorological and marine observations at Horns Rev; and (4) the Danish Meteorological Institute and the Royal Danish Administration of Navigation and Hydrography for the tidal observations in Port of Esbjerg. In addition, we thank Dr Henning Skriver, Denmark Technical University for comments on SAR theory and Dr Yaron Levy for comments on the manuscript and language.