MODELING THERMAL INFRARED (2-14-MU-M) REFLECTANCE SPECTRA OF FROST AND SNOW

Existing theories of radiative transfer in close-packed media assume t hat each particle scatters independently of its neighbors. For opaque particles, such as are common in the thermal infrared, this assumption is not valid, and these radiative transfer theories will not be accur ate. A new method is proposed, called ''diffraction subtraction'', whi ch modifies the scattering cross section of close-packed large, opaque spheres to account for the effect df close packing on the diffraction cross section of a scattering particle. This method predicts the ther mal infrared reflectance of coarse (>50 mu m radius), disaggregated gr anular snow. However, such coarse snow is typically old and metamorpho sed, with adjacent grains welded together. The reflectance of such a w elded block can be described as partly Fresnel in nature and cannot be predicted using Mie inputs to radiative transfer theory. Owing to the high absorption coefficient of ice in the thermal infrared, a rough s urface reflectance model can he used to calculate reflectance from suc h a block. For very small (<50 mu m), disaggregated particles, it is i ncorrect in principle to treat diffraction independently of reflection and refraction, and the theory fails. However, for particles larger t han 50 mu m, independent scattering is a valid assumption, and standar d radiative transfer theory works.