Carbon nanotube forests are vertically grown tubular formations of graphene. Samples were grown with an injection chemical vapor deposition method on substrates of silicon with various deposited layers and bare fused silica. The morphology of the forest is characterized by the height, density, and presence of defects. Total diffuse reflectance and transmittance measurements were taken in the 2-16 �m spectral range and correlated to the forest’s specific morphology. From these correlations, the conditions necessary to maximize the absorption of the forest were found and exploited to cater sample growth for specific substrates to make ideal absorbers. From the transmittance data, the absorption coefficient is found via Beer-Lambert’s Law and also correlated to sample morphology, giving us an indication of the height of the forest needed for ideal absorption. Two models were used to attempt to reproduce the experimental absorption coefficient: an effective medium theory using a Maxwell Garnett approximation and by treating the carbon nanotube forest as an effective cylindrical waveguide with walls of graphite. Each model leads to a set of fitting parameters providing a better physical understanding of the forests. It was found that the effective medium theory gave results loosely corroborated with electron microscopy, but had trouble fitting the experimental data, and the index of refraction it provides does not behave like a unified medium. The waveguide model fits the data well, but it requires more experimental evidence to be more conclusive. The theoretical models need more work, but fabrication of ideal absorbers has been achieved on various substrates providing framework for their usage in radiometry and spectroscopy.