Spectral subtraction is a powerful and conventional tool for the elimination of background interferences, but this technique has been suffered from severe limitations in the case of atmospheric water vapor. In order to understand the real reasons of the problems above , IR spectra of water vapor at different humidity have been investigated. The following conclusions have been reached: (1) The spectral features of water vapor will change as the relative humidity of air changes. Namely, if the spectrum of atmospheric air with x% relative humidity is given, we cannot predict and obtain the spectrum of air with y% humidity regardless of how carefully the scaling factor has been chosen. In general, spectral subtraction is not a good practice to remove water vapor interferences. (2) The relative composition of water clusters (H2O)n changes with the relative humidity, and this is the main reason of spectral subtraction limitation. (3) The spectrum of water vapor between relative humidity x% and y% can be matched near perfectly by a combination of x% spectrum and y% spectrum. This provides a new and effective way for the elimination of water vapor interferences. (4) The high efficiency of the humidity titration method has been demonstrated experimentally.