Ocorrência e estrutura vertical de florestas com castanheiras na Amazônia brasileira

Abstract

The Brazil nut tree is a symbol of the theory about the importance of the moderate use and management of the forest resources by the populations that inhabited and inhabit the Amazon region, for the conservation of its biodiversity. The factors that explain the occurrence of brazil nut tree and its relations with the mother forest, are not yet completely elucidated. To help clarify this issue, we tested the hypothesis that, depending on the scale of analysis, Brazil nut forests present a own vertical structure pattern and that their potential distribution can be modeled with the MaxEnt algorithm. One of the main approaches of this study, besides the focus on the Brazil nut tree itself, was the validation of advanced methods of data collection and analysis. The use of these methods, such as terrestrial LiDAR and MaxEnt modeling, are fundamental for carrying out this research. The objective of the first chapter was to evaluate if forests with Brazil nut present a vertical structure pattern, also verifying if the abundance of Brazil nut can be associated with attributes of the forest canopy. The study was carried out in two scales: 1) regional, covering forests in four states of the Brazilian Legal Amazon; 2), in two areas with different forest typologies in southern Amapá. Terrestrial LiDAR was used to collect the data. The forests in the Eastern Amazon presented greater structure of the vegetation and the larger Brazil nut trees. It was identified that the forest in Amazonas presented a smaller structure in relation to the other states. In the Amazon there is a pattern of distribution of leaf density along the vertical profile of forests with Brazil nut trees, having a higher concentration in the lower strata. The stratum (8 m to 18 m) had a higher proportion of leaf density in all forests. In the local scale, inverse results were found, with differences along the profile and similarities in height structure. This indicates that the spatial scale of analysis is important and that variations in the vertical profile of forests with Amazonian Brazil nut trees are more dependent on the forest typology than on wide geographic variations. The objective of the second chapter was to test if the MaxEnt algorithm can be used to model the occurrence of Brazil nuts in Amapá and whether the results are differentiated according to the selection method of topobioclimatic variables and the scale of analysis. The modeling was carried out in two scales: 1) State, State of Amapá; 2) Location in Resex Cajari. Three methods (PCA, expert panel, pre-modeling) were used to select 9 variables, 3 of each group: climate, soil and topography. The results indicate that the variables selected as the most important in the modeling vary between methods and scales. However, all adjusted models were significant. In general, the modeling with MaxEnt was effective to model the regions with real occurrence of Brazil nut trees on both scales, probably due to the high density of presence points its that entered in the modeling. However, new tests and approaches are still needed, which consider a greater rarefaction of occurrence points, absence information and interactions with biotic factors, to try to detect the aggregate pattern of occurrence in southern Amapá and to improve the prediction of potential areas suitable for the establishment of Brazil nut trees. The methods used in this research proved to be efficient for studies of the structure of the vertical profile of the Amazon forest and the modeling of the distribution of Brazil nut trees in Amapá