Spectral reflectance of sea foam in the visible and near-infrared: In situ measurements and remote sensing implications

The spectral reflectance of sea foam was measured at the Scripps Institution of Oceanography Pier, La Jolla, California, by viewing the sea surface radiometrically in a region of breaking waves. Foam reflectance was found to decrease substantially with wavelength in the near-infrared, contrary to the findings of previous studies, theoretical as well as experimental. Values in the visible (0.44 mu m) were reduced by typically 40% at 0.85 mu m, 50% at 1.02 mu m, and 85% at 1.65 mu m. The spectral effect was explained by the nature of the foam, which is composed of large bubbles of air separated by a thin layer of water (foam stricto sensu) and of bubbles of air injected in the underlayer. The presence of bubbles in the underlayer enhances water absorption and thus reduces reflectance in the near-infrared. For ocean color remote sensing, affected by the presence of foam and aerosols, the consequences of neglecting the spectral dependence of foam are dramatic. With only a small amount of foam, in the presence of aerosols or not and thus irrespective of aerosol type, the errors in the retrieved water reflectance at 0.44 mu m are above 0.01, which does not meet the accuracy goal of 0.001 for biological applications. Since under normal conditions the effect of foam may have the same magnitude as the effect of aerosols, atmospheric corrections will be inaccurate (and useless) in many cases, even taking into account the spectral dependence of the foam reflectance. Space observations potentially contaminated by an effective foam reflectance (product of reflectance and fractional coverage) above 0.001, i.e., corresponding to wind speeds above 8 m s(-1), should be eliminated systematically. Utilization of near-infrared wavelengths above 0.9 mu m for atmospheric corrections of ocean color, possible with the moderate-resolution imaging spectrometer (MODIS), would aggravate the problem. The measurements also indicated that foam significantly affects the retrieval of aerosol turbidity at 0.85 and 1.02 mu m for wind speeds above 10 m s(-1) but impacts minimally turbidity estimates at 1.65 mu m. Over the oceans the spectral range above 1 mu m is definitely recommended for remote sensing of tropospheric aerosol load and type from space.