Modelos Volumétricos Mistos Em Clones De Eucalyptus No Polo Gesseiro Do Araripe, Pernambuco

Abstract

ResumoO presente estudo teve como objetivo estimar o volume de clones de Eucalyptus utilizando modelos mistos. A base de dados foi proveniente de um povoamento de clones de Eucalyptus localizado na Chapada do Araripe, no semiárido do estado de Pernambuco. Foram cubadas rigorosamente, pelo método de Smalian, 89 árvores na idade de sete anos e meio. O modelo de Schumacher e Hall foi utilizado como testemunha, para comparação com os modelos não lineares mistos. Os ajustes dos modelos mistos foram realizados adotando-se seis estruturas distintas para a matriz de variância e covariância. A seleção da melhor equação se deu por meio do critério de informação de Akaike (AIC), Teste da Razão de Máxima Verossimilhança (TRMV), Erro Percentual Absoluto Médio (EPAM) e Teste de Vuong. De acordo com os critérios adotados, os modelos mistos obtiveram melhores ajustes quando comparados com o modelo clássico de Schumacher e Hall, apresentando uma redução no erro percentual absoluto médio de 4,6% para 3,2%. Diante dos resultados obtidos, os modelos mistos não lineares se mostram bastante eficazes para modelagem do volume de Eucalyptus e tendem a contribuir para a redução dos custos do inventário com uma maior acurácia. Palavras-chave: Inventário florestal; estimação volumétrica; comparação entre modelos. AbstractMixed volumetric models in clones of Eucalyptus in the gypsum pole of Araripe, Pernambuco. The present research aimed to estimate the volume of Eucalyptus clones using mixed models. The database was derived from an experiment of Eucalyptus clones, located in the Chapada do Araripe, semiarid of the Pernambuco State. 89 trees were cubed by the method Smalian at the age of seven and a half years. The model of Schumacher and Hall was used as a control for comparison with the nonlinear mixed models. The adjustments of the mixed models were performed by adopting six distinct structures for the matrix of variances and covariances. The selection the best equation was done using the Akaike information criterion (AIC) Test Likelihood Ratio (TRMV), Mean Absolute Percentage Error (EPAM) and Vuong test. According to the criteria adopted, mixed models obtain better adjustments when compared with the classical model of Schumacher and Hall, resulting in a decrease in mean absolute percentage error of 4.6% to 3.2%. Based on these results, the nonlinear mixed models appear quite effective for modeling the volume of Eucalyptus and tend to contribute to the reduction of inventory costs with greater accuracy. Keywords: Forest inventory; volumetric estimation; comparing models. INTRODUÇÃOCom o aumento do desenvolvimento dos setores sociais e industriais, tem crescido a busca por recursos energéticos, que tem sido suprida principalmente pelas florestas nativas, com a procura da biomassa, que possui a vantagem de ser de fácil aquisição e de menor custo. Dessa forma, cresce o número de áreas desertificadas em ritmo acelerado, em razão da exploração irresponsável dos recursos florestais (OLIVEIRA et al., 2006;BARROS et al., 2010).

Estimation of the Basic Wood Density of Native Species Using Mixed Linear Models

“…The attainment of higher results, in terms of accuracy, in all models with random effects, in comparison to the regression model with only fixed effects, was validated from the analysis of statistical indicators [ 2 R , RMSE (%) and B (%)]. The above observation can be attributed to the ability of mixed linear models to extract the relationships of independent variables with clustered data in a better way (species) as compared to traditional regression models (Calegario et al, 2005;Gouveia et al, 2015).…”

“…It takes into consideration several assumptions, such as the independence between observations and homogeneity of variance. But these assumptions are not always fulfilled (Gouveia et al, 2015).…”

“…Thus, as an alternative, more accurate statistical techniques such as mixed models are used, which allows generalization of spatiotemporal correlations structures and non-constant variance. Moreover, in comparison to traditional regression models, the mixed models are capable of extracting the relationships of independent variables with clustered data (Calegario et al, 2005;Fausto et al, 2008;Gouveia et al, 2015).…”

Assessing Alternatives to Estimate the Stem Volume of a Seasonal Semi-Deciduous Forest

“…Considering that there are different alternatives that may be used to estimate the volume of trees in even-aged stands, in contrast to classical regression models, such as hybrid models (GOUVEIA et al, 2015), artificial neural networks (BINOTI et al, 2014;GÖRGENS et al, 2014) and support vector machine (BINOTI et al, 2016), we verified the possibility of using these approaches on the estimation of volumes for the different species that form the uneven-aged forest. For this purpose, in addition to the input (or explanatory) variables, such as tree diameter and height, the species can also be included as random effects in a hybrid structure model by including random intercepts and inclinations, or as categorical variables in the artificial neural networks (ANN), and support vector machines (SVM).…”

Performance of maize genotypes of Rio Grande do Sul using mixed models

“…The use of mixed models is already performed frequently in perennial plants such as Brazilian Chestnut (Camargo et al, 2010), Aroeira (Carderalli et al, 2013), Eucalyptus (Gouveia et al, 2015;de Mendonça et al, 2015), Coffee (Carias et al, 2014) and fruitful (Della Bruna et al, 2012). However in annual plants, especially in the case of maize crop, publications addressing this theory are more recent, especially in Brazil (Souza et al, 2015).…”