Fires have an obvious impact on the ecology of a region and are an important component of land management in certain regions. Approaches to detecting fires using satellite observations have concentrated on using multi-spectral imagery at visible, 3.9, 10.7, and 12 micron wavelengths to identify and catalogue fire. In this study, we demonstrate and validate an approach to fire detection that uses high-spectral resolution measurements around 2400 cm-1 (4.2 microns) combined with imager data. The fire detection approach is demonstrated and the fire detection feature explained using radiative transfer calculations. The fraction of the S-HIS field of view containing fire, hot air over a cooler surface, and the undisturbed background was estimated by fitting the shape and amplitude of the observed fire-signature to that determined from calculation.