Dinâmica e modelagem temporal de vegetação campestre sob distúrbios

Abstract

The development of economically viable production systems, which can make compatible good beef cattle profit and the preservation of natural grassland ecosystems, should be supported by knowledge of dynamic of vegetation that occur after some management practices. Vegetation dynamic studies show the existence of populations replacement processes which can be observed in space and time, in different scales. Modeling these systems would be one alternative to achieve this premise. To achieve this purpose, this thesis, is divided in two chapters. In the are presented, respectively, temporal dynamics of grasslands vegetation, based on species composition (taxons) or Plants´ Functional Types (PFT) groups. These were based on the traits: Specific Leaf Area (SLA) e Leaf Dry Matter Content (LDM), which grouped Poaceae family species in four groups (A, B, C and D). Data used in this thesis came from vegetation relévés from 1995 to 2001, and data predict by Markov chains. Typical disturbance factors of grassland vegetation, grazing and burning, and their interactions, combined with relief position, express the scenarios which should be modeled along the time. Results obtained by using ordination and randomization analysis show that is possible using this tool to build up predictive models of vegetation composition in grassland areas, independently from the evaluation being based on taxons or PFT. When it was based in taxons, a certain degree of imprecision in the estimation of forage mass of species with a low initial contribution. Grazing and relief position were determinant of ordination trajectory, and that was not possible to identify in burning. When based in PFTs, results obtained from model validation, from ordination analysis and from randomization tests, showed a tendency of trajectories and of predicted values of functional groups, which is more suitable for plots under grazing. Using this model, as a management tool for areas under grazing, is in dependence of new validation researches, using new data, but for this is necessary to make strong the exchange of informations among research groups in this area.