Biomass models are useful for several purposes, especially for quantifying carbon stocks and dynamics in forests. Selecting appropriate equations from a fitted model is a process which can involves several criteria, some widely used and others used to a lesser extent. This study analyzes six selection criteria for models fitted to six sets of individual biomass collected from woody indigenous species of the Tropical Atlantic Rain Forest in Brazil. Six models were examined and the respective fitted equations evaluated by the residual sum of squares, adjusted coefficient of determination, absolute and relative estimates of the standard error of estimate, and Akaike and Schwartz (Bayesian) information criteria. The aim of this study was to analyze the numeric behavior of these model selection criteria and discuss the ease of interpretation of them. The importance of residual analysis in model selection is stressed. The adjusted coefficient of determination ( $$ R^{2}_{adj.} $$ R a d j . 2 ) and the standard error of estimate in percentage (Syx%) are relative model selection criteria and are not affected by sample size and scale of the response variable. The sum of squared residuals (SSR), the absolute standard error of estimate (Syx), the Akaike information criterion and the Schwartz information criterion, in turn, depend on these quantities. The best fit model was always the same within a given data set regardless the model selection criteria considered (except for SSR in two cases), indicating they tend to converge to a common result. However, such criteria are not always closely related across different data sets. General model selection criteria are indicative of the average goodness of fit, but do not capture bias and outlier effects. Graphical residual analysis is a useful tool to this detection and must always be used in model selection. It is concluded that the criteria for model selection tend to lead to a common result, regardless their mathematical formulation and statistical significance. Relative measures of goodness of fitting are easier to interpret than the absolute ones. Careful graphical residual analysis must always be used to confirm the performance of the models.

The traditional method used to estimate tree biomass is allometry. In this method, models are tested and equations fitted by regression usually applying ordinary least squares, though other analogous methods are also used for this purpose. Due to the nature of tree biomass data, the assumptions of regression are not always accomplished, bringing uncertainties to the inferences. This article demonstrates that the Data Mining (DM) technique can be used as an alternative to traditional regression approach to estimate tree biomass in the Atlantic Forest, providing better results than allometry, and demonstrating simplicity, versatility and flexibility to apply to a wide range of conditions.

Este trabalho teve como objetivo avaliar o desempenho das funções densidade de probabilidade (fdp) Weibull 3 Parâmetros (3P) e Beta ajustadas pelo método da Máxima Verossimilhança para descrever a distribuição diamétrica de espécies arbóreas de um fragmento de Floresta Ombrófila Mista. A área de estudo faz parte da Floresta Nacional de Irati (FLONA), onde foram instaladas 25 parcelas permanentes (100 m x 100 m) com área total amostrada de 25 ha e, para os ajustes, foram utilizados dados de todas as árvores com diâmetro à altura do peito (DAP) superior ou igual a 10 cm. Os ajustes foram feitos para as 20 espécies com maior valor de importância (VI) com dados agrupados em classes de 5 cm de DAP. A programação não linear (PNL) foi utilizada com o objetivo de minimizar os desvios entre valores observados e estimados. Para avaliação dos ajustes, foram utilizados os testes de aderência de Kolmogorov-Smirnov e de Hollander-Proschan, além do Erro-Padrão de Estimativa (%) e Índice de Reynolds. Os resultados indicaram que as funções Weibull 3P e Beta apresentaram resultados satisfatórios, que ambas poderiam ser recomendadas para avaliar a distribuição diamétrica das espécies consideradas e que o uso da PNL nem sempre melhorou os ajustes.

Phytosociological analysis in native forests is performed considering the horizontal and vertical structure of the studied population, whose most expressive parameters are the density, dominance, frequency, value of coverage and value of importance of species. Many phytosociological studies include a value of importance of species calculated by adding density, dominance and frequency, in their relative forms, however, this estimator is a mathematical impropriety because the result is a sum of indexes and does not express the true meaning of species' value. This is because the estimator is still influenced predominantly by the density of occurrence of species and cannot capture the relevant hierarchical participation of emerging species that occur with low density in the biocoenosis. In this work, we propose a new index to characterize the value of importance of species based on the hierarchy and spatial absolute probability of the species in the biocoenosis, illustrated with data from a forest fragment of semideciduous tropical forest located in Cassia, MG, Brazil. The new index appropriately expressed the importance of species in the evaluated fragment.

O objetivo desta pesquisa foi avaliar o comportamento de sete funções densidade de probabilidade no ajuste da distribuição diamétrica de Eschweilera ovata em um fragmento de Floresta Ombrófila Densa no estado de Pernambuco, Brasil. Os dados utilizados são provenientes dos diâmetros de 274 árvores dessa espécie medidos durante a realização de um estudo sobre a estrutura do componente arbóreo e classificação sucessional do presente fragmento. Foram testadas as seguintes funções probabilísticas: Beta, Gama, Log-Normal, Normal, SB de Johnson, Weber e Weibull 3P, utilizando intervalo de classe de 4 cm. Os resultados do teste de Kolmogorov-Smirnov indicaram que a função mais eficiente foi a Log-Normal, seguida de Weber, Weibull 3P, Beta, SB de Johnson e Gama. A distribuição Normal foi rejeitada pelo teste de Kolmogorov-Smirnov. O estudo também identificou que a distribuição em diâmetro de Escheweilera ovata nesse fragmento é do tipo decrescente. Palavras-chave: Distribuições probabilísticas; floresta nativa; “J” invertido. Abstract Diametric distribution of Escheweilera ovata in a Dense Ombrophilus Forest fragment - Igarassu, Pernambuco, Brazil. This research aimed to evaluate the behavior of seven probability density functions for fitting the diameter distribution of Escheweilera ovata in a Dense Ombrophilous Forest fragment in the State of Pernambuco, Brazil. The data used consisted of diameters of 274 trees of this species measured during a research improved to evaluate the tree components and succession classification of this fragment. The following probabilistic functions were tested: Beta, Gamma, Log-Normal, Normal, Johnson SB , Weber and Weibull 3P, it was used diametric class interval of 4 cm. The results of the Kolmogorov - Smirnov test indicated that the most efficient function was the Log-Normal, followed by Weber, Weibull 3P, Johnson SB , Beta, and Gama. The Normal distribution was rejected by the Kolmogorov - Smirnov test. This research also indicated that the Escheweilera ovata diameter distribution in this fragment is decreasingly. Keywords: Probabilistic distributions; native forest; inverted “J”.