Extinction and transformation

Abstract

The basic principles of heterochromatic extinction show that the approach used in the visible should not work well in the infrared, where molecular line absorption rather than continuous scattering dominates the extinction. Not only does this extinction change very rapidly with wavelength (so that stellar color becomes only weakly correlated with effective extinction), but also many of the lines are saturated (so that Forbes's curve-of-growth effect is much more severe in the IR.) Furthermore, broadband IR colors are more under-sampled than those in the visible, so aliasing errors make them correlate even less with extinction, and enhance the difficulties of transformation to a standard system. Reduction to outside the atmosphere is difficult, but a rational approximation for the Forbes effect may help. Plausible assumptions about the probability distribution function of line strengths, and band-model approaches, may be useful. The only solution to the transformation problem is to satisfy the sampling theorem, which may be difficult in the IR because of gaps due to saturated telluric absorptions.