Relações entre ambiente, traços, composição e funcionamento de comunidades vegetais de cerrado

• Autor(a) principal: Carvalho, Gustavo Henrique de
• Data de Publicação: 2013
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/1796
• topic: Cerrados; Comunidades vegetais; Funcionamento de comunidades; Modelos de equações estruturais; Maxent; Savanna; Species pool; Cerrado; Traits; CIENCIAS BIOLOGICAS::ECOLOGIA

Identifying the relative importance of deterministic and stochastic factors to the assembly of local communities has been tagged as one of the fundamental questions ecologists need to answer. Niche-based theories predict that the subset of species from the regional pool that occur in a local community presents a trait distribution that has been shaped by biotic and abiotic factors. The neutral theory, on the other hand, predicts that the species that compose the regional pool are functionally equivalent. Local communities formed from stochastic processes would, thus, present abundances not significantly different from those in the species pool. We collected traits on all woody individuals in 100 25m2 quadrats to test if we could predict local abundances using a pool of species and traits. We combined plots into large scale, intermediate scale, and fine scale samples. We used random sampling, spatial distance, soil characteristics, and fire to assemble species in different scales and environmental gradients. In the Brazilian cerrado, fire and nutrient-poor soils are likely promoters of habitat filtering. To test if traits improved the predictions generated by the information present in the pool, we used maximum entropy models coupled with permutation tests. We could accurately predict local abundances of the 73 species in the pool. Dispersal limitation was the main factor assembling communities at all the scales we studied, but the effects of stochasticity became more important as the scale became local. Traits explained little of the uncertainty present in local abundances, but coupled with pool frequencies they yielded large coefficients of determination.