Accurate radiative transfer simulations of signals received by sensors deployed on satellite platforms for remote sensing purposes can be computationally demanding depending on channel width and the spectral variation of atmospheric and surface optical properties. Therefore, methods that can speed up such simulations are desirable. While it is common practice to use atmospheric "window" channels to minimize the influence of gaseous absorption, the impact of the underlying surface as well as clouds and aerosols has received less attention. To reduce the number of monochromatic computations required to obtain a desired accuracy, one may average the inherent optical properties (IOPs) over a spectral band to generate effective or mean IOP values to be used in "quasi-monochromatic" radiative transfer computations. Comparison of radiances produced by computations based on mean (quasi-monochromatic) IOPs with benchmark results in typical shortwave infrared window channels, revealed that while this approach may be sufficient for gaseous absorption, it led to significant errors in the presence spectrally varying surface IOPs, in general, and snow/ice surfaces, in particular. To solve this problem, a new method was developed in which a satellite channel is represented by a few subbands. This new method significantly reduces the error resulting from IOP averaging to be typically less than 1%. An additional correction was also developed to further reduce the error incurred by use of mean gas IOPs for large solar zenith angles to be less than 0.01%.