The objective of this study was to obtain neural networks that would precisely estimate inside-bark diameter (d ib ) and heartwood diameter (d h ) and compare to the results obtained by the Taper models. The databank was formed so as to eliminate inconsistent and biased data, and stratified: minimum d ib of 4, 6 and 8 cm and minimum d h of 10, 15 and 20 cm. The adjusted Taper model used was the Kozak model. For the fitting of artificial neural networks (ANN), tests were performed to identify the independent variables and the database scope level, i.e., the following input variables were tested: diameter at breast height (dbh), total height (H), height at diameter d ib or d h (h) and outside-bark diameter at h (d ob ), bark thickness at 1.3 m and project, and the scope at database level or project level. The estimates obtained by the neural networks and Kozak model were evaluated by residual graphs in function of the respective diameter observed and graph of the observed versus estimated values. ANN were found to be more efficient in estimating inside-bark and heartwood diameters for Tectona grandis trees than the Kozak model. The variables that must be used to fit the networks are dbh, H, h and d ob . Stratification by project results in precision gain, with precision being higher for wider commercial diameters. Thus, linear-type artificial neural networks can be efficient in describing the taper of Tectona grandis trees.

ABSTRACT:The objective of this study was to evaluate the effectiveness of fatigue life, Frechet, Gamma, Generalized Gamma, Generalized Logistic, Log-logistic, Nakagami, Beta, Burr, Dagum, Weibull and Hyperbolic distributions in describing diameter distribution in teak stands subjected to thinning at different ages. Data used in this study originated from 238 rectangular permanent plots 490 m 2 in size, installed in stands of Tectona grandis L. f. in Mato Grosso state, Brazil. The plots were measured at ages 34, 43, 55, 68, 81, 82, 92, 104, 105, 120, 134 and 145 months on average. Thinning was done in two occasions: the first was systematic at age 81months, with a basal area intensity of 36%, while the second was selective at age 104 months on average and removed poorer trees, reducing basal area by 30%. Fittings were assessed by the Kolmogorov-Smirnov goodness-of-fit test. The Log-logistic (3P), Burr (3P), Hyperbolic (3P), Burr (4P), Weibull (3P), Hyperbolic (2P), Fatigue Life (3P) and Nakagami functions provided more satisfactory values for the k-s test than the more commonly used Weibull function. 34, 43, 55, 68, 81, 82, 92, 104, 105, 120, 134

RESUMO -As funções de densidade probabilidade Weibull e Hiperbólica foram comparadas quanto a eficiência de descrever a estrutura diamétrica de povoamentos de Teca (Tectona grandis L. f.) submetidas a desbaste. As duas funções com três e quatro parâmetros, foram ajustadas com dados de 98 parcelas permanentes, retangulares (490 m²), instaladas em um povoamento desbastado de Tectona grandis, no Estado do Mato Grosso e medidas durante 10 anos. Os ajustes foram feitos por Máxima Verossimilhança e a aderência foi avaliada pelo teste Kolmogorov-Smirnorv (α = 1%). Também foram comparadas a soma de quadrados dos resíduos (SQR) dos diferentes ajustamentos. Todas as funções apresentaram aderência aos dados pelo teste de Kolmogorov-Smirnorv (α = 1%). A função hiperbólica apresentou menor soma de quadrados de resíduos e menores valores para o teste de aderência. Foi possível concluir que a função hiperbólica foi mais eficiente para descrever a estrutura diametrica dos povoamentos estudados. EFFICIENCY OF THE WEIBULL AND HYPERBOLIC FUNCTIONS FOR DESCRIBING THE DIAMETRIC DISTRIBUTIONS OF Tectona grandis STANDS ABSTRACT -The Weibull probability density functions and

The Guazuma crinita Mart. is a dominant species of great economic importance for the inhabitants of the Peruvian Amazon, standing out for its rapid growth and being harvested at an early age. Understanding its vertical growth is a challenge that researchers have continued to study using different hypsometric modeling techniques. Currently, machine learning techniques, especially artificial neural networks, have revolutionized modeling for forest management, obtaining more accurate predictions; it is because we understand that it is of the utmost importance to adapt, evaluate and apply these methods in this species for large areas. The objective of this study was to build and evaluate the efficiency of the use of a deep neural network for the prediction of the total height of Guazuma crinita Mart. from a large-scale continuous forest inventory. To do this, we explore different configurations of the hidden layer hyperparameters and define the variables according to the function HT = f(x) where HT is the total height as the output variable and x is the input variable(s). Under this criterion, we established three HT relationships: based on the diameter at breast height (DBH), (i) HT = f(DBH); based on DBH and Age, (ii) HT = f(DBH, Age) and based on DBH, Age and Agroclimatic variables, (iii) HT = f(DBH, Age, Agroclimatology), respectively. In total, 24 different configuration models were established for each function, concluding that the deep artificial neural network technique presents a satisfactory performance for the predictions of the total height of Guazuma crinita Mart. for modeling large areas, being the function based on DBH, Age and agroclimatic variables, with a performance validation of RMSE = 0.70, MAE = 0.50, bias% = −0.09 and VAR = 0.49, showed better accuracy than the others.

Optimizing tree spacing in a forest plantation is one of the main management techniques to improve stand quality and productivity. Its influence on growth from an early age is an important matter for forest management. This study aims to evaluate the effect of tree spacing on early growth rate and yield over time in Eucalyptus grandis × Eucalyptus camaldulensis hybrids. The data were obtained from an experiment in Itamarandiba, Minas Gerais, Brazil. The plots were composed of five planting spacing (3.00 m × 0.50 m, 3.00 m × 1.00 m, 3.00 m × 1.50 m, 3.00 m × 2.00 m, and 3.00 m × 3.00 m) measured at the ages of 7, 12, 24, 36, 48, 61, 77, 85, and 102 months. Growth and yield were analyzed by fitting the Gompertz model and a baseline exponential model up to 36 months of age to evaluate the influence of early growth on the harvest age. A Pearson correlation matrix was also generated to find out the relationship between the mean annual increment in the respective treatments during the studied period. It was observed that a positive correlation in the average annual increase in the 3 × 2 and 3 × 3 spacings. It was verified that tree spacing influenced the yielded wood volume and the optimal harvest age. The early growth rate influences the optimal harvest, which may explain a possible loss of yield during the productive cycle of the forest stand.