On sea surface properties and characteristics in the infrared.

摘要

This study addressed issues relevant to IR remote sensing of SST using a combination of new instrumentation, development of analysis techniques and high quality laboratory and field measurements. The data used were high spectral resolution infrared interferometric measurements of the sea surface. Using spectral statistics techniques, the effective incidence angle of ship-borne radiometric measurements, the sea surface spectral emissivity in the 8–12μm wavelength region and the depth of the oceanic molecular boundary layer were determined and characterized.; The slope of the sea surface depends on the wave field and the effective incidence angle, in turn, depends on the surface slope. The effective incidence angle of ship-borne measurements was found to be 3–4° from instrument pointing angle at wind speeds up to 15ms−1. For pointing angles of 55° or greater, particularly at 8–12μm, this offset can result in SST errors of up to 0.6K. The effect was skewed towards smaller values of incidence angle at higher wind speeds due to the tilted facets of the wave field.; Analysis techniques were developed and evaluated to calculate the spectral emissivity of fresh and seawater without a priori knowledge of the water's refractive index. Emissivity measured over a wide range of wind speeds (0 to 13ms−1) increased with wind speed at a 55° incidence angle and was constant at 40°. The use of wind speed dependent models of sea surface emissivity which predict the opposite trend, can bias SST retrievals by up to 0.4K at a wind speed of 10ms −1 and incidence angles greater than 40°.; In turbulent free convection, microscale thermal profiles showed a linear gradient down to about 80μm depth. The gradients were not as steep as predicted due to convective motions in the sublayer. For a linear subsurface gradient, the skin depths were 0.1mm-1mm and exhibited a linear decrease with net water-air heat flux. The oceanic skin layer depth was estimated to be 0.11mm, an order of magnitude smaller than the laboratory measurements. It had the largest variance at low wind speeds and heat fluxes, decreasing rapidly to about 30μm ± 20μm with wind speeds over 3ms−1.