Anisotropy in reflected GPS measurements of ocean winds

The probability density function (PDF) of the ocean surface slope can be estimated from the code-correlation waveform of reflected GPS signals. Anisotropy in this PDF is found to correspond to the local near surface wind direction, suggesting the ability to resolve this direction using scattered signals from two or more GPS satellites. A two-stage estimation process was applied to sets of waveform data collected from an airborne delay-mapping GPS receiver. First, an isotropic normal distribution was assumed for the PDF. The mean square slope (MSS) was fit to the measured waveform data for each satellite independently. Differences between the MSS estimates from two satellites at different azimuths were observed. In the second step, a bidirectional normal PDF was used with MSS constrained to that obtained in the first step, and an assumed value was given for the ratio of upwind to crosswind slopes. The direction of the principal axes was then varied to minimize the total residuals for both satellites. The results were compared with buoy recordings of the local wind direction.