Modelagem de árvore individual para povoamentos de Khaya ivorensis A. Chev. no Brasil

Abstract

Khaya ivorensis A. Chev., known as African mahogany, is a species for the noble uses with quality wood. The species has good adaptation in Brazil, but still has few areas of commercial plantation and advanced ages, when compared to other exotic species, and consequently, few works on a forest modeling and even on a silviculture of this species in Brazil. It is necessary to invest in studies that return future information of the growth and yield of African mahogany and, from these studies, to improve the forest's execution of the species. This work had as main objective the individual tree independent of the distance modeling of African mahogany in Brazilian plantations with different characteristics. The data used comes from continuous forest inventories on ten farms with plantations of Khaya ivorensis in the Brazilian states of Minas Gerais, Goiás and Pará. The plantations have an arrangement ranging from 4x5 m to 12 x 12 m and ages from 1.1 to 15 years. This study is structured in four chapters. The first chapter consists of a background of the topics worked in the research and more detailed description of the data. The second chapter discusses distanceindependent competition indices for African mahogany, an essential variable in individual modeling. The best performing index was the Basal Area Larger (BAL), which is a competition index independent of the widely used distance in the literature. The third chapter refers to the modeling of annual increment in diameter independent of distance with mixed effect. The results were compared with the studies of Ribeiro (2017) and Oliveira (2017) developed for the species with diametric and global class models and with SisMogno (software developed for the prognosis of African mahogany). The equation developed presented results with biological realism and statically coherent. And, finally, the fourth chapter deals with the concept and modeling of bole height. The logistic model with the variables: total height, natural logarithm of basal area, natural logarithm of diameter at breast height and height ratio, site index and age was selected to predict bole height.