Spectral sea surface reflectance of skylight

In examining the dependence of the sea surface reflectance of skylight rho(s) on sky conditions, wind speed, solar zenith angle, and viewing geometry, Mobley [Appl. Opt.38, 7442 (1999).] assumed rho(s) is independent of wavelength. Lee et al. [Opt. Express18, 26313 (2010).] showed experimentally that rho(s) does vary spectrally due to the spectral difference of sky radiance coming from different directions, which was ignored in Mobley's study. We simulated rho(s) from 350 nm to 1000 nm by explicitly accounting for spectral variations of skylight distribution and Fresnel reflectance. Furthermore, we separated sun glint from sky glint because of significant differences in magnitude, spectrum and polarization state between direct sun light and skylight light. The results confirm that spectral variation of rho(s)(lambda) mainly arises from the spectral distribution of skylight and would vary from slightly blueish due to normal dispersion of the refractive index of water, to neutral and then to reddish with increasing wind speeds and decreasing solar zenith angles. Polarization moderately increases sky glint by 8 - 20% at 400 nm but only by 0 - 10% at 1000 nm. Sun glint is inherently reddish and becomes significant (>10% of sky glint) when the sun is at the zenith with moderate winds or when the sea is roughened (wind speeds > 10 m s(-1)) with solar zenith angles < 20. We recommend a two-step procedure by first correcting the glint due to direct sun light, which is unpolarized, followed by removing the glint due to diffused and polarized skylight. The simulated rho(s)(lambda) as a function of wind speeds, sun angles and aerosol concentrations for currently recommended sensor-sun geometry, i.e., zenith angle = 40 and azimuthal angle relative to the sun = 45, is available upon request. (C) 2017 Optical Society of America