Numerical classification and digital elevation model to spatial characterization of soil attributes

Abstract

One of needs of modern agriculture is the prediction of spatial variability of soil properties at more detailed scales for sustainable management and optimization of management practices. The mathematical model associated with knowledge of variability of soil attributes and mapping of relief forms has helped in agricultural planning. In this regard the aim of this study was to characterize the spatial variability of physical and chemical properties of Oxisols and Ultisols using numerical classification and the digital elevation model. Two distinct landforms: convex for the Oxisol (158 ha) and linear for the Ultisol (172 ha). 53 samples from the Oxisol and 57 samples from the Ultisol were taken. Multivariate analysis of clusters of attributes studied from their euclidean distances was performed. This analysis by dendograms along with digital elevation models for different soils characterized was more homogeneous in Ultisol groups, and less homogeneous for the Oxisol in convex landform. These quantitative methods showed that the landforms conditioned the spatial pattern of soil attributes.