A method for satellite imagery (Landsat TM/ETM+, IRS, etc.) classification based on spectral space modeling is described. The 3D model is built in xyz-axes, where x and y are the first two principal components of the image matrix in logarithmic form (bands G, R, NIR, SWIR2), and z is the moisture stress index MSI (SWIR1/NIR). Ambiguous spectral classes are decomposed using the geomorphometric model and time series imagery. The method is optimal for taiga ecosystem structure and dynamics modeling.