An alteration zone may be clearly expressed in an image of a spectral band ratio, as long as the spatial variation of its tonality is strong enough. A measure of the spatial variation of the target of interest (in our case the alteration zone) can be the semivariogram of the band ratio.In the present paper, the spatial variation of the band ratio, in terms of the statistical behavior of the individual bands, is studied. The semivariograms of the individual bands is simulated by the stable semivariogram model, which is quite flexible in describing the spatial variation of the tonality. The correlation between the spectral bands is also taken into account. Using proper distributions for the reflectance of individual bands and taking into account theorems of statistics, a mathematical expression for the semivariogram of the spectral band is derived. This expression provides information about how the spatial variation of the band ratio is affected by the statistical parameters of the individual bands. Experimentation with satellite images shows that theoretical predictions agree with real data.The results and conclusions of this paper may help in assessing the efficiency of the images of spectral band ratios over potential mineralization zones.