Topographic effects on the texture of high-resolution forest-stand images measured by the semivariogram

By studying the statistical relationship between terrain gradient and the r ange of the semivariogram of simulated high resolution images of forest sta nds, we assessed the effects of topography on estimates of tree size and de nsity obtained through texture measures. Three-dimensional computer models of hardwood and softwood forest stands were overlaid on slopes of varying g radient. Using a geometrical-optical approach, one-meter-resolution images were generated in 120 series representing different combinations of forest types and sun-terrain geometry. The range of the semivariogram of these ima ges was measured in four directions and its relation with gradient was eval uated through regression. Results show that topography affects texture most ly in sparse stands, and that the gradient-induced absolute error in the es timates of tree size and density is low.