Afilamento e otimização de sortimentos de Tectona grandis Linn.f. para a maximização do volume e da renda

Abstract

The study was conducted in order to describe the shape of bole and evaluate the income from assortments of trees coming from Tectona grandis L.f plantations located in the central-south region of Mato Grosso state, and to develop a simulator to provide the forest producer technical and economic information to the appropriate planning of forest stand. The database consisted of T. grandis trees divided into ten classes of diameter at breast height (DBH) covered by different procedures and measurement positions. Randomly approximately 20 trees per DBH class comprised the setting database, making a total of 148 trees. Nonsegmented, segmented and variable models were adjusted and the accuracy of the equations was evaluated by graphical analysis of the residuals, adjusted coefficient of determination, standard error of the estimate in percentage and graphical analysis of the average profiles. The best equation was submitted to stratifications of DBH, artificial form factor and "r" shape parameters, and modifications based on the study of the shape change points of the trees. Mixed modeling was applied to the selected equation, testing random effects for DBH class and for trees. DBH classes were defined with class centers of 10 to 60 cm and intervals of 2 cm for the elaboration of the assortment table. The Forest Assortment Analysis, Simulation and Optimization System was developed with Mixed Integer Linear Programming with two main modules, Log Trace Module and Assortment Optimization Module, and a complementary module for the purpose of estimating dendrometric parameters, managing the exchanging information and presenting the results. The scenarios were constructed in the order to simulate the assortment optimization for data from the diameters distribution for three occurrences of thinning and for three different sites. With the Schöepfer equation, three points of change of the T. grandis bole were determined in different positions when compared to that established in the model of Clark III et al. The equation of Clark III et al. is the most adequate to estimate the diameters along the trunk for the species and the adjustment with mixed effect increases the flexibility of the model, and this is more accurate in relation to the tree than in relation to the DBH class. The equation of Clark III et al. with mixed effect is accurate to estimate the profile, especially at the ends of the trunks, and the integral of the function allows estimating the trunk volume and its assortment. The auxiliary functions developed to operate the assortment optimization system enable both data entry and assortment from individual tree data, DBH classes and even stocking. The T. grandis assortment optimization system for volume and income maximization allows the evaluation of several scenarios for the technical and economic planning of forest production in the south-central region of Mato Grosso.