Validation of RadarSat SAR-derived Winds Using Atmospheric Model Estimates and Remotely Sensed Measurements

To validate SAR-derived winds from RADARSAT, we constructed high quality, vector wind maps by combining remotely sensed wind data with wind estimates from a numerical weather prediction (NWP) model. The remotely sensed data is from the NSCAT and ERS-2 scatterometers and the TOPEX-Poseidon and ERS-2 altimeters. NWP model wind fields are produced using the MC2 atmospheric model (Canadian mesoscale compressible community atmospheric model). The grid for these fields is about 25km in the Northwest Atlantic. Numerical model MC2 outputs are blended with remotely sensed winds, using optimal interpolation (OI). Correlation functions and variances for the optimal interpolation are derived from remotely sensed data. This study focuses on the Labrador Sea Deep Convection Experiment (LSDCE) data, during 1 January - 31 March 1997. Verification of OI winds is achieved by comparison with independent meteorological buoy data. These OI winds are then the basis to validate vector wind fields constructed from RADARSAT SAR measurements, using the Vachon-Dobson algorithm.