DETERMINAÇÃO DA TAXA DE INFILTRAÇÃO DE ÁGUA NO SOLO E LÂMINA INFILTRADA EM LATOSSOLO VERMELHO-AMARELO

Abstract

The aim of this study was to evaluate the efficiency of using an artificial neural network (ANN) to estimate the stem taper of eucalypt trees in a silvopastoral system composed with two spatial arrangements. The data used were collected out of 35 sample-trees scaled in a silvipastoral system with spatial arrangements of 12 m x 4 m and 12 m x 2 m. The taper model proposed by Garay was fitted for each spatial arrangement. Also, the ANN with Multilayer Perceptron configuration, using the spatial arrangement as a categorical variable, was trained. The others input variables for the ANN were the diameter at breast height - 1.30 m height, total height, height of each section and the corresponding diameters. The accuracy of the methods was evaluated using the Root Mean-square Error, the correlation between observed and estimated diameters, dispersion of percentage errors and the mean absolut deviaton. The ANN achieved a similar performance compared to the two functions of tapering, proving to be an appropriate methodology for small eucalyptus plantations in silvopastoral system, where there may be restrictions for logging.