Avaliação da influência de borda em vegetação heterogênea com análise de padrão espacial

• Ano de defesa: 2015
• Autor(a) principal: Dodonov, Pavel
• Orientador(a): Matos, Dalva Maria da Silva
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ecologia e Recursos Naturais - PPGERN
• País: BR
• Palavras-chave em Português: Conservação; Cerrados; Influência de borda; Cadeias de Markov; Transformada de Wavelet; Diversidade estrutural; Efeitos de borda; Gramíneas invasoras; Modelos autoregressivos; Padrão espacial; Tundra
• Área do conhecimento CNPq: CIENCIAS BIOLOGICAS::ECOLOGIA
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/1858

Edge influence, the set of modifications in the natural environment next to abrupt anthropogenic or natural edges, is a serious threat to biodiversity in fragmented environments. Common examples of edge influence include an increased abundance of exotic species and modifications in vegetation structure and/or composition. Edges of linear disturbances, such as roads, powerlines, and firebreaks, are especially important due to their ubiquity in the landscape. Although such disturbances are much narrower than other anthropogenic land uses, they have been shown to affect adjacent plant and animal communities. The combination of edge influence with the natural variation in vegetation may lead to complex and not easily discernible patterns. However, this natural variation is not usually taken into account in studies of edge influence. We employed a spatial pattern approach to study edge influence from narrow linear disturbances while accounting for variation in vegetation (both natural and from disturbance history), in Brazilian cerrado and in Canadian forest-tundra ecotone. Our main objectives were 1) to assess the overall heterogeneity in the cover of native and invasive graminoids, structural diversity and species composition, and 2) to verify whether the spatial pattern of different response variables is affected by anthropogenic edges. Secondary objectives include: 3) to explore different ways in which wavelet transforms and Monte Carlo simulations may be used to assess edge influence, 4) to compare different structural diversity indices, and 5) to assess whether the spatial patterns of different plant species are related to their functional traits. We sampled five 300-1350 m-long transects, three in Brazil and two in Canada. The transects contained one to five anthropogenic edges and one to ten natural edges and were divided into contiguous 1x1 m quadrats. In each quadrat, we estimated the cover of native and invasive graminoids (two transects in Brazil), sampled different structural elements to measure the structural diversity (two transects in Brazil and the two Canadian transects), and identified plant species and their dispersal syndromes and lifeforms (one transect in Brazil). We analyzed the data by means of continuous and discrete wavelet transforms and multiresolution analysis, and assessed significance by means of full randomizations, Markov chain simulations, and autoregressive models. Although some edge-related patterns were apparent for most variables, they were not observed at all edges, and similar patterns were also observed far from edges. The different structural diversity indices showed similar patterns, and species traits, namely lifeform and 8 dispersal syndrome, were not related to the spatial distribution of different species. The main conclusion of our study is that although narrow linear disturbance edges may alter vegetation structure and species composition, the conditions at edges are not necessarily different from what may be observed due to the natural variation in plant communities in non-forest vegetation. Keywords: autoregressive models, cerrado, dispersal syndromes, edge effects, invasive grasses, Markov chain, spatial pattern, structural diversity, tundra, wavelet transform.