Retrieval of Wind Wave Elevation Spectra From Sunglint Data

A new expression for the mean value of ocean surface sunglint, modeled as a binary-valued random process, is calculated. Multiple sunglint realizations are generated by applying the specular condition for particular viewing geometries to ocean surface elevations with Gaussian roughness spectra. The sunglint mean value is used to determine the relationship between the second-order statistics of the surface slope and sunglint random processes, and this relationship is successfully inverted to retrieve the slope autocorrelation and power spectrum from the simulated sunglint data. The inversion model is then applied to an image recorded coincident with the NASA AIRSAR PACRIM2 field campaign, to retrieve the 1-D elevation power spectrum.