Long-term Variability of Ocean Surface Winds

Just a few decades ago, almost all ocean wind measurements came from merchant ships. However, the quality and geographical distribution of these wind reports were uneven. The quality of historical wind-stress have been well investigated [e.g., Cardone et al., 1990; Trenberth, et al, 1990; Kent et al., 1999]. Today, operational numerical weather prediction (NWP) also gives us wind information, but NWP depends on models, which are limited by our knowledge of the physical processes and the availability of data. Spacebased sensors have been shown to provide superior high wavenumber wind forcing of global oceans [Liu et al.,1998]. The European Remote Sensing (ERS) satellites have provided over ten years of ocean surface wind vectors [e.g. Quilfen et al., 200]. Considerable amount of science application have made use of the almost five years of data from QuikSCAT at much higher spatial and temporal resolutions [see Liu 2002 for a review]. The Special Sensor Microwave/ Imager (SSMI) have been in operation since 1978, but only wind speeds are retrieved from the radiance. The SSMI wind speeds have been combined with other directional information to provide fourteen years of ocean surface wind vectors, whose application in forcing the ocean have been demonstrated [e.g., Busalacchi et al., 1990; Liu et al.,1996]. This paper intends to examine the longterm and low-frequency variability of the ocean surface wind-field derived from these spacebased sensors .