Relações espaciais e ambientais da biodiversidade em florestas tropicais

Abstract

This study aimed to evaluate issues related to spatial ecology and how spatial analysis can help ecologists to understand patterns of species distribution. Initially, we reviewed the current literature on spatial ecology and illustrated basic concepts with simulated data represented various situations that ecologists frequently face. In a second step we evaluated the importance of clearly defining the spatial variables to be included in ecological models, assessing the differences between analyses using a matrix that defines dispersal routes through stream distance and routes of dispersal in a straight line overland. One of most frequently used techniques spatial ecology, spatial eigenvector functions, is generally considered to be so flexible in generating spatial patterns that it should generate the same patterns no matter what distance matrix is used. Using data from fish, aquatic insects (Trichoptera) and simulated data, we show that the technique of eigenvectors is not as flexible as expected. Therefore, the correct definition of the matrix of distances is important for an effective analysis. In the next step, we evaluated two approaches to analyze ecological data. One of them uses raw data and the other using distance matrices. A recent discussion on the use of these two approaches has been polemical and we try to clarify what types of questions each of these approaches is better able to analyze. To do this we used data on the distribution of Trichoptera in 92 streams of central Amazonia, sampled in three distinct regions (Ducke Reserve, at PDBFF areas, and at the municipality of Presidente Figueiredo). We also evaluated the effect of spatial extent and environmental heterogeneity to on the distribution of Trichoptera species. We found that environmental factors were more related to the distribution of Trichoptera than undefined factors that caused spatial clumping at all scales examined. We analyzed the distribution of anuran species at 72 sample plots in Ducke Reserve and observed that the species distribution pattern is much related to the type of reproduction of the species. The distribution of species with aquatic reproduction is better explained by environmental patterns, while species with terrestrial reproduction are better explained by spatial patterns. We conclude that anurans with aquatic reproduction are better indicated for biomonitoring and for studies evaluating the consequences of environmental disturbances than the use of species with terrestrial reproduction or than the use of both. Finally, we evaluated the congruence in the distribution patterns of 22 taxa (15 plants and 7 animal groups) sampled at Ducke reserve. We also evaluated if it is possible to reduce the taxonomic resolution of species to genera and the numeric resolution from abundance data to presence-absence data with little loss of information. Presence-absence data and genus-level identification was sufficient to capture most of the spatial patterns of most groups. There was strong congruence in the distribution pattern of plant groups, while animal groups were less concordant. Measured environmental factors were closely related to the high congruence among the groups, indicating that environment is primarily responsible for the distribution of plants in the Ducke Reserve. Although some groups showed strong spatial patterns, there was no relationship between the congruence among groups and spatial factors. The groups that were more concordant with other groups, and possible candidates as surrogates, were the lianas of the family Bignoniaceae, herbs, ferns, and trees of the families Fabaceae and Lecythidaceae.