RESUMEN
In tropical forests, the spatial distribution of trees may present random, uniform, or grouped patterns that can simultaneously be affected by site and species characteristics. In Central Amazon, topographic gradients and soil water levels drive differences in tree species distribution and in forest dynamics at local scales. Knowing this kind of information can be useful for a forest manager to plan harvesting operations considering the microhabitat preference of merchantable species to reduce the disturbances caused by logging activities. Thus, the spatial variation of tree species is an important information to be considered to support the planning process of forest logging. The present study aims to evaluate the spatial distribution pattern of six species and analyze the relationship between the topography and the population densities and stem size of those species. The study was carried out in a forest production compartment managed by a private company located in the municipality of Silves, state of Amazonas, Brazil. The spatial pattern of the six species was characterized by Ripley's K function. Spatial distribution of diameter at breast height (DBH) and tree density based on kernel incidence calculation were evaluated for topographic classes of slope, elevation, and distance from streams, which were mapped using geographic information systems (GIS). The means of DBH and density of each species were compared among topographic classes by ANOVA and Tukey's test. The results demonstrated the predominance of the aggregate distribution pattern for the six species up to 1105 m (p < 0.01). The tree species Minquartia guianensis Aubl., Protium puncticulatum J.F.Macbr, Manilkara elata (Allemão ex Miq.) Monach, and Caryocar glabrum Aubl. Pers showed an increase in the tree density as the distance from the streams and elevation increased, standing spatially incident on plateaus. Kernel densities of Dinizia excelsa Ducke and Goupia glabra Aubl. were higher closer to streams. The DBH averages followed similar trends of population density for M. guianensis, M. elata, and C. glabrum, and the opposite pattern for D. excelsa, which presented larger individuals in less densely populated areas. P. puncticulatum and G. glabra mean DBH distribution was not affected by the topographic variables analyzed. Topography-related variables showed effects on variations of density and tree size, suggesting that species may be spatially sensitive to the habitat variability available in the study area. In view of logging planning, spatial distribution must be considered in decisions related to cutting down trees and maintenance of remaining trees, especially because some species are more aggregated in smaller scales. Moreover, as topographic variations affect the spatial distribution of size and density, the timber yield will vary spatially in the area, bringing implications for planning logging intensities, roads, skid trails and forest operations. Finally, the procedures and information generated in this study can be reproduced and applied to other species and managed areas to support the planning toward minimizing impacts on the spatial structure of commercial species, as well as to increase the chances of future stock recovery of managed forests in the Amazon.