Modelagem da distribuição diamétrica de espécies amazônicas na Flona do Jamari, RO

Abstract

The management of natural forests must be carried out in such a way that the exploited species are able to recover the stock that was removed during exploitation, and so that the continuity of forest production is guaranteed. For this to happen efficiently, many studies must be carried out and the analysis of the diameter distribution is a fundamental tool in planning, as it allows knowing the structure of the forest and the species in it. In uneven natural forests, the diameter distribution should behave in a decreasing way, to indicate a balanced forest with a high potential for self-regeneration. The probabilistic density functions are currently the best way to describe the diametric structure of forest stands, and several of them have been used in studies of the forest area. In this context, this study aimed to evaluate the performance of probabilistic density functions to express the diameter distribution of a forest fragment in the Jamari National Forest in Rondônia and of five commercial species exploited through a Forest Management Plan in that area. The study also aimed to evaluate whether logging caused changes in the shape of the diameter distribution of species and the forest as a whole, in addition to testing the adherence of the best model selected in the adjustment carried out prior to logging, for modeling the distribution after logging. The database used was provided by the company Amata, referring to the forest inventory of the area under forest concession in 2014. Six probabilistic density functions widely used in the forestry sector were adjusted, namely: Normal, Log-Normal, Gamma, Beta, Negative Exponential and Weibull 3 parameters; and tested the adherence of the models by the Kolmogorov-Smirnov test (K-S) at 1% of significance. The best fitted models for the fragment and for the forest species were selected based on the lowest D value calculated for the K-S test. The best model for the forest fragment was the Weibull 3P however it did not show adherence to the observed frequencies, the worst model was the negative exponential. For the species Peltogyne paniculata, the fdp Beta had better results, however it also did not show adherence to the real frequencies. The Gamma model was the best fit for the species Dinizia excelsa, and the Normal model was the only one to show no adherence for this species. For the species Astronium lecointei, the fdp's of Weibull 3P and Beta had similar results for the adhesion test, being the second best for this species. Couratari stellata had the Log-Normal and Weibull 3P models as the best to estimate the diameter frequencies of this species, with the former showing better results. The species Dipteryx odorata was the one that presented the best results for the adherence test having high values of significance, the Weibull 3P model was the best placed followed by the beta model that presented similar results for the K-S test. All frequencies estimated by the models after exploration with the initial parameters and the new adjusted parameters were statistically equal to each other, not differing in the use of the results from a statistical point of view.