Comparison of four ERS-1 scatterometer wind retrieval algorithms with buoy measurements

Wind velocity retrievals from the ERS-1 scatterometer are compared with extensive high-quality hourly buoy winds using four different algorithms in two oceanic regions during 1994. The retrieved winds exhibit significantly different wind velocity biases. The retrieved direction exhibits a clockwise offset of about 10 degrees for three of the four algorithms in the equatorial Pacific, whereas less offset was observed at midlatitudes. All the retrievals exhibit 180 degrees ambiguities in direction about as often as the true directions. The retrieved wind speed consistently overestimates the lightest winds and underestimates the highest winds. Two of the algorithms (ESA and OSU/JPL) significantly underestimate the highest buoy winds with the largest bias of about 20% in the equatorial region. The underlying geophysical models employed in these two algorithms were primarily tuned using surface wind (weather) analysis fields. The remaining two algorithms (IFREMER and CU) were tuned using buoy measurements resulting in smaller biases in both regions.